{"id": "c-darwin-01-dessert-01", "uid": "c000", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "prompt_sha256_by_order": ["ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d"], "request_sha256": "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "compiled_prompt_sha256_by_order": ["2ca65940a5765f77af59cf411c0b326a074bbc7804d92041b9a96860a2ba21f0", "48dbee6ba046a0f5d1492a021834e80774f5e6d672ab585b30be1393344d7180", "8469ceacf69d457d0646180caed55919f309bd99e78602af7c4c31eca9d50741", "218d7301ee961a1bc56ec764ba057f602e68d84c1990b15df5371a4b82fa233e", "f13adde1c9aaa8a22e12b94a59507cb89bd9e3cf532778ae7d31880236e63212", "3c31ef585bd4ffd3fc8a990fd1ab1073f330e1883dd469671aa08c5027d3084d"], "compiled_tokens_by_order": [279, 279, 279, 279, 279, 279], "choice": "ibn_sina", "probs": {"darwin": 0.2953069693020607, "einstein": 0.02139196510676243, "hypatia": 0.02139196510676243, "ibn_sina": 0.42966910501141903, "sagan": 0.2299852989388882, "socrates": 0.00225469653410722}, "confidence": 0.3082722408814944, "seconds": 0.17970754916677834, "seconds_all_orders": 1.07824529500067, "seconds_by_order": [0.21237184699930367, 0.1781775320014276, 0.1722510190011235, 0.17212180999922566, 0.17132512599710026, 0.17199796100248932], "argmax_stable": false, "spread": {"darwin": 0.5139179355144309, "einstein": 0.22905432878471882, "hypatia": 0.036830283934572605, "ibn_sina": 0.3953453463983043, "sagan": 0.2767857738699453, "socrates": 0.017927810239148426}, "spread_2dp": {"darwin": 0.51, "einstein": 0.22, "hypatia": 0.03, "ibn_sina": 0.4, "sagan": 0.27, "socrates": 0.02}, "choices_by_order": ["ibn_sina", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.2953069693020607, "einstein": 0.02139196510676243, "hypatia": 0.02139196510676243, "ibn_sina": 0.42966910501141903, "sagan": 0.2299852989388882, "socrates": 0.00225469653410722}, {"ibn_sina": 0.06042235089254327, "hypatia": 0.005620154105177389, "sagan": 0.30685001704386666, "socrates": 0.006779199888072471, "darwin": 0.5732713038802123, "einstein": 0.0470569741901279}, {"hypatia": 0.016053298461629114, "darwin": 0.46914614986751163, "sagan": 0.32243911899000344, "ibn_sina": 0.1344126223655317, "einstein": 0.0560315172295802, "socrates": 0.001917293085743929}, {"ibn_sina": 0.03432375861311472, "einstein": 0.015231075556441953, "darwin": 0.7812051452500255, "sagan": 0.15382841382931828, "hypatia": 0.0143082713384206, "socrates": 0.0011033354126789815}, {"einstein": 0.05173201740011982, "sagan": 0.045653345120058135, "socrates": 0.019031145651827407, "ibn_sina": 0.06642522519828635, "darwin": 0.8092249048164916, "hypatia": 0.007933361813216636}, {"socrates": 0.015616643423546166, "ibn_sina": 0.07930785925898998, "hypatia": 0.042450438039749995, "einstein": 0.24428540434116078, "sagan": 0.2155811126776999, "darwin": 0.4027585422588532}], "prompt_tokens": 279, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.2953069693020607, "einstein": 0.02139196510676243, "hypatia": 0.02139196510676243, "ibn_sina": 0.42966910501141903, "sagan": 0.2299852989388882, "socrates": 0.00225469653410722}, "confidence": 0.3082722408814944, "choice": "ibn_sina"}}, "usage": {"input_tokens": 279, "output_tokens": 1}, "latency_ms": 212.30403700246825, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 279, "compiled_prompt_sha256": "2ca65940a5765f77af59cf411c0b326a074bbc7804d92041b9a96860a2ba21f0", "compiled_ids_sha256": "39875e59687c95458e3c8bef14fa71362355147df3bbbb6895cd7ad2bb135fb8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.875, 16.25, 16.25, 19.25, 18.625, 14.0], "runtime_probabilities": [0.2953069508075714, 0.021391963586211205, 0.021391963586211205, 0.4296690821647644, 0.2299852967262268, 0.0022546963300555944], "decide_seconds": 0.2053407019993756, "probabilities_f64_minus_runtime_max_abs": 2.2846654623709384e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.06042235089254327, "hypatia": 0.005620154105177389, "sagan": 0.30685001704386666, "socrates": 0.006779199888072471, "darwin": 0.5732713038802123, "einstein": 0.0470569741901279}, "confidence": 0.4096048545322374, "choice": "darwin"}}, "usage": {"input_tokens": 279, "output_tokens": 1}, "latency_ms": 178.1527930033917, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 279, "compiled_prompt_sha256": "48dbee6ba046a0f5d1492a021834e80774f5e6d672ab585b30be1393344d7180", "compiled_ids_sha256": "1c43b23c4ff7ee2b0043a4d253a4bbf977882d588e01e9dacd72a1c7b6474d65", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.25, 14.875, 18.875, 15.0625, 19.5, 17.0], "runtime_probabilities": [0.060422345995903015, 0.005620154086500406, 0.30685001611709595, 0.0067791990004479885, 0.5732712745666504, 0.04705697298049927], "decide_seconds": 0.1742231839998567, "probabilities_f64_minus_runtime_max_abs": 2.9313561911159525e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.016053298461629114, "darwin": 0.46914614986751163, "sagan": 0.32243911899000344, "ibn_sina": 0.1344126223655317, "einstein": 0.0560315172295802, "socrates": 0.001917293085743929}, "confidence": 0.3137676256327536, "choice": "darwin"}}, "usage": {"input_tokens": 279, "output_tokens": 1}, "latency_ms": 172.22043899892014, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 279, "compiled_prompt_sha256": "8469ceacf69d457d0646180caed55919f309bd99e78602af7c4c31eca9d50741", "compiled_ids_sha256": "b850e58e668e249c6ab16ac15630067755df76d9a0cbf85e7e4b392802f09665", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 19.375, 19.0, 18.125, 17.25, 13.875], "runtime_probabilities": [0.016053300350904465, 0.469146192073822, 0.32243916392326355, 0.13441264629364014, 0.05603152513504028, 0.001917293295264244], "decide_seconds": 0.16842247900058283, "probabilities_f64_minus_runtime_max_abs": 4.49332601126784e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.03432375861311472, "einstein": 0.015231075556441953, "darwin": 0.7812051452500255, "sagan": 0.15382841382931828, "hypatia": 0.0143082713384206, "socrates": 0.0011033354126789815}, "confidence": 0.5933625532253219, "choice": "darwin"}}, "usage": {"input_tokens": 279, "output_tokens": 1}, "latency_ms": 172.09977000311483, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 279, "compiled_prompt_sha256": "218d7301ee961a1bc56ec764ba057f602e68d84c1990b15df5371a4b82fa233e", "compiled_ids_sha256": "03f544c4c81ba324ffa4af0f748ddef4ec3b4480551eb9123c4a9b91dd0308a9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 15.9375, 19.875, 18.25, 15.875, 13.3125], "runtime_probabilities": [0.034323763102293015, 0.015231077559292316, 0.7812051773071289, 0.15382841229438782, 0.014308272860944271, 0.0011033355258405209], "decide_seconds": 0.16861485700064804, "probabilities_f64_minus_runtime_max_abs": 3.205710341624979e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.05173201740011982, "sagan": 0.045653345120058135, "socrates": 0.019031145651827407, "ibn_sina": 0.06642522519828635, "darwin": 0.8092249048164916, "hypatia": 0.007933361813216636}, "confidence": 0.5762174093312682, "choice": "darwin"}}, "usage": {"input_tokens": 279, "output_tokens": 1}, "latency_ms": 171.302906001074, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 279, "compiled_prompt_sha256": "f13adde1c9aaa8a22e12b94a59507cb89bd9e3cf532778ae7d31880236e63212", "compiled_ids_sha256": "26ae9000f113b6924b052671c5bf5b7966533710e6a39fc0ca6ea6923d555ae4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 17.0, 16.125, 17.375, 19.875, 15.25], "runtime_probabilities": [0.05173201486468315, 0.04565334692597389, 0.019031144678592682, 0.06642522662878036, 0.8092249035835266, 0.007933362387120724], "decide_seconds": 0.16784813300182577, "probabilities_f64_minus_runtime_max_abs": 2.5354366661733962e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.015616643423546166, "ibn_sina": 0.07930785925898998, "hypatia": 0.042450438039749995, "einstein": 0.24428540434116078, "sagan": 0.2155811126776999, "darwin": 0.4027585422588532}, "confidence": 0.19551601781270045, "choice": "darwin"}}, "usage": {"input_tokens": 279, "output_tokens": 1}, "latency_ms": 171.9757020000543, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 279, "compiled_prompt_sha256": "3c31ef585bd4ffd3fc8a990fd1ab1073f330e1883dd469671aa08c5027d3084d", "compiled_ids_sha256": "ef6a744ad93767063b46b3134aab606d4679224932cb3b5550b206fd09ee53ce", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.875, 17.5, 16.875, 18.625, 18.5, 19.125], "runtime_probabilities": [0.015616643242537975, 0.07930785417556763, 0.0424504354596138, 0.24428540468215942, 0.21558111906051636, 0.40275853872299194], "decide_seconds": 0.1685447080017184, "probabilities_f64_minus_runtime_max_abs": 6.382816469807651e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:32Z"} {"id": "c-darwin-01-dessert-01", "uid": "c001", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "prompt_sha256_by_order": ["ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66"], "request_sha256": "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "compiled_prompt_sha256_by_order": ["dfdf84665a6deae6826cfe5cb74cbdf7164c3c10210051c62958077957da1655", "1b455f62cd412deb3a24d61fab9d3554f21bb2aed9d2368dd8596679da0f9778", "5868f7f26b62319d9696c2266c1d8d6f001d23651b72baa28b1386b803b7ce58", "fd499f127c17859447da2316773685477d9baac7cb8267200f21b6bfeca90015", "c104604437579117b904b7c741fcf93ef03da60367be03d47a5c692e96a90bde", "7cb668279fdd9d43553280b896c143d250d5ea31cae48d76db701d614598e31a"], "compiled_tokens_by_order": [379, 379, 379, 379, 379, 379], "choice": "darwin", "probs": {"darwin": 0.9920496923803619, "einstein": 0.0009046322234155576, "hypatia": 0.002170099890878105, "ibn_sina": 0.00203862018510099, "sagan": 0.00203862018510099, "socrates": 0.0007983351351424453}, "confidence": 0.9673380100087969, "seconds": 0.19422366783389103, "seconds_all_orders": 1.1653420070033462, "seconds_by_order": [0.19412331800049287, 0.19442136699944967, 0.19438235699999495, 0.19399178899766412, 0.1941856280027423, 0.1942375480030023], "argmax_stable": true, "spread": {"darwin": 0.021837557173342015, "einstein": 0.011308536139906015, "hypatia": 0.005715314265916217, "ibn_sina": 0.003985257705744851, "sagan": 0.0015754389692282725, "socrates": 0.004292870895465449}, "spread_2dp": {"darwin": 0.020000000000000018, "einstein": 0.01, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.0, "socrates": 0.01}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.9920496923803619, "einstein": 0.0009046322234155576, "hypatia": 0.002170099890878105, "ibn_sina": 0.00203862018510099, "sagan": 0.00203862018510099, "socrates": 0.0007983351351424453}, {"ibn_sina": 0.005858207735173561, "hypatia": 0.0033379061512077346, "sagan": 0.0016784046171306954, "socrates": 0.0020245424199508842, "darwin": 0.9851990574810152, "einstein": 0.0019018815955218664}, {"hypatia": 0.002615314401744721, "darwin": 0.9911682722826294, "sagan": 0.0011605387307905286, "ibn_sina": 0.0021681717922429275, "einstein": 0.0017974775489867723, "socrates": 0.0010902252436055943}, {"ibn_sina": 0.0029497179534889377, "einstein": 0.0019044810500480301, "darwin": 0.9865456082627101, "sagan": 0.0021580597557570103, "hypatia": 0.004863262732483263, "socrates": 0.0015788702455126014}, {"einstein": 0.004547101255798734, "sagan": 0.001149687243599407, "socrates": 0.0017806704368399232, "ibn_sina": 0.003125173942687436, "darwin": 0.9819004645606119, "hypatia": 0.007496902560462638}, {"socrates": 0.005091206030607894, "ibn_sina": 0.0018729500294287092, "hypatia": 0.007885414156794323, "einstein": 0.012213168363321572, "sagan": 0.0027251262128276795, "darwin": 0.9702121352070199}], "prompt_tokens": 379, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.9920496923803619, "einstein": 0.0009046322234155576, "hypatia": 0.002170099890878105, "ibn_sina": 0.00203862018510099, "sagan": 0.00203862018510099, "socrates": 0.0007983351351424453}, "confidence": 0.9673380100087969, "choice": "darwin"}}, "usage": {"input_tokens": 379, "output_tokens": 1}, "latency_ms": 194.0941480024776, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 379, "compiled_prompt_sha256": "dfdf84665a6deae6826cfe5cb74cbdf7164c3c10210051c62958077957da1655", "compiled_ids_sha256": "935f8d246f67d04de94e49c25e2c59f166760a7b2d6dddf4f9fb2ac6ee141755", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.875, 13.875, 14.75, 14.6875, 14.6875, 13.75], "runtime_probabilities": [0.9920497536659241, 0.0009046322666108608, 0.0021701001096516848, 0.0020386201795190573, 0.0020386201795190573, 0.000798335182480514], "decide_seconds": 0.18957887400028994, "probabilities_f64_minus_runtime_max_abs": 6.12855621806574e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.005858207735173561, "hypatia": 0.0033379061512077346, "sagan": 0.0016784046171306954, "socrates": 0.0020245424199508842, "darwin": 0.9851990574810152, "einstein": 0.0019018815955218664}, "confidence": 0.9447288139829846, "choice": "darwin"}}, "usage": {"input_tokens": 379, "output_tokens": 1}, "latency_ms": 194.39603700084263, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 379, "compiled_prompt_sha256": "1b455f62cd412deb3a24d61fab9d3554f21bb2aed9d2368dd8596679da0f9778", "compiled_ids_sha256": "abf2faa76564d81fb58d134901e31cc8f085074c776142296365eda850d32e3c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.25, 14.6875, 14.0, 14.1875, 20.375, 14.125], "runtime_probabilities": [0.005858208052814007, 0.003337906440719962, 0.001678404863923788, 0.0020245425403118134, 0.9851991534233093, 0.0019018817692995071], "decide_seconds": 0.18985576200066134, "probabilities_f64_minus_runtime_max_abs": 9.59422941093635e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.002615314401744721, "darwin": 0.9911682722826294, "sagan": 0.0011605387307905286, "ibn_sina": 0.0021681717922429275, "einstein": 0.0017974775489867723, "socrates": 0.0010902252436055943}, "confidence": 0.9641208465174013, "choice": "darwin"}}, "usage": {"input_tokens": 379, "output_tokens": 1}, "latency_ms": 194.35761700151488, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 379, "compiled_prompt_sha256": "5868f7f26b62319d9696c2266c1d8d6f001d23651b72baa28b1386b803b7ce58", "compiled_ids_sha256": "a514ce09513e6a1603befcb5277126f6b04f732acfadf3dd2792edbd5f03f2e9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.1875, 21.125, 14.375, 15.0, 14.8125, 14.3125], "runtime_probabilities": [0.002615314442664385, 0.9911682605743408, 0.0011605386389419436, 0.0021681718062609434, 0.0017974774818867445, 0.001090225181542337], "decide_seconds": 0.18995414199889638, "probabilities_f64_minus_runtime_max_abs": 1.1708288560896563e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0029497179534889377, "einstein": 0.0019044810500480301, "darwin": 0.9865456082627101, "sagan": 0.0021580597557570103, "hypatia": 0.004863262732483263, "socrates": 0.0015788702455126014}, "confidence": 0.9487586685556257, "choice": "darwin"}}, "usage": {"input_tokens": 379, "output_tokens": 1}, "latency_ms": 193.96777899964945, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 379, "compiled_prompt_sha256": "fd499f127c17859447da2316773685477d9baac7cb8267200f21b6bfeca90015", "compiled_ids_sha256": "7642123f71f1bc74d7c3ca6c07e85bb7c21a09942791cc6afdd5183de9fa4dc6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.8125, 14.375, 20.625, 14.5, 15.3125, 14.1875], "runtime_probabilities": [0.0029497179202735424, 0.001904481090605259, 0.9865455627441406, 0.002158059738576412, 0.004863262642174959, 0.0015788702294230461], "decide_seconds": 0.18960146400058875, "probabilities_f64_minus_runtime_max_abs": 4.5518569469216175e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.004547101255798734, "sagan": 0.001149687243599407, "socrates": 0.0017806704368399232, "ibn_sina": 0.003125173942687436, "darwin": 0.9819004645606119, "hypatia": 0.007496902560462638}, "confidence": 0.9351341359661447, "choice": "darwin"}}, "usage": {"input_tokens": 379, "output_tokens": 1}, "latency_ms": 194.16213799922843, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 379, "compiled_prompt_sha256": "c104604437579117b904b7c741fcf93ef03da60367be03d47a5c692e96a90bde", "compiled_ids_sha256": "d0da3fb5c81cae0502389616395e7847afb7792ce34455d715ace63571161b0d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0, 13.625, 14.0625, 14.625, 20.375, 15.5], "runtime_probabilities": [0.004547101445496082, 0.0011496871011331677, 0.0017806702526286244, 0.003125173971056938, 0.9819003939628601, 0.007496901787817478], "decide_seconds": 0.18974554300075397, "probabilities_f64_minus_runtime_max_abs": 7.05977517556633e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.005091206030607894, "ibn_sina": 0.0018729500294287092, "hypatia": 0.007885414156794323, "einstein": 0.012213168363321572, "sagan": 0.0027251262128276795, "darwin": 0.9702121352070199}, "confidence": 0.901735366050308, "choice": "darwin"}}, "usage": {"input_tokens": 379, "output_tokens": 1}, "latency_ms": 194.21315799991135, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 379, "compiled_prompt_sha256": "7cb668279fdd9d43553280b896c143d250d5ea31cae48d76db701d614598e31a", "compiled_ids_sha256": "86a7c6efd2f4acf88c8f5375d7ed27508013e0000fea6565e7ae9495cf7c34c4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.25, 14.25, 15.6875, 16.125, 14.625, 20.5], "runtime_probabilities": [0.005091205704957247, 0.0018729500006884336, 0.00788541417568922, 0.012213167734444141, 0.0027251262217760086, 0.9702121019363403], "decide_seconds": 0.18980919199748314, "probabilities_f64_minus_runtime_max_abs": 3.3270679544727955e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:34Z"} {"id": "c-darwin-01-dessert-01", "uid": "c002", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "prompt_sha256_by_order": ["6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc"], "request_sha256": "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "compiled_prompt_sha256_by_order": ["36256d9ffe4d50c3d22a8cefe2d22c38863f669314f6c78f6e3165edd736e631", "e42bd9b0cedfb08ec693aa27f8b2df6de90ca7d55e8463294125dc41439e0a37", "0d3acb839327b5ed8b9a89af5a322a76c70bad66bcaf5f2e160d4bb6c9d57850", "9b6d91cb0b3666a5057a4df12fc8247dda9ac305984b9d9ef3535630cdbab79e", "c1bb3d6c6a862369b6084f743e6e387dfa5e1afab61d21235ee28083ce7263db", "4d89b0a4d92b78766d8c6df2302113378d8eabe3158ff0cc3f5de3002ee99f8f"], "compiled_tokens_by_order": [296, 296, 296, 296, 296, 296], "choice": "darwin", "probs": {"darwin": 0.3512688666047678, "einstein": 0.11404030424940888, "hypatia": 0.04753907155416017, "ibn_sina": 0.24142332599051308, "sagan": 0.21305533739829882, "socrates": 0.032673094202851166}, "confidence": 0.13814500572755695, "seconds": 0.1737173491668121, "seconds_all_orders": 1.0423040950008726, "seconds_by_order": [0.17456421099996078, 0.17355452899937518, 0.17358211900136666, 0.17330203000165056, 0.17316252199816518, 0.17413868400035426], "argmax_stable": false, "spread": {"darwin": 0.5640152706109218, "einstein": 0.4370770463384639, "hypatia": 0.04033701972874178, "ibn_sina": 0.20033974956441675, "sagan": 0.1527532957598908, "socrates": 0.07985802647524404}, "spread_2dp": {"darwin": 0.56, "einstein": 0.44, "hypatia": 0.04, "ibn_sina": 0.19999999999999998, "sagan": 0.15, "socrates": 0.08}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.3512688666047678, "einstein": 0.11404030424940888, "hypatia": 0.04753907155416017, "ibn_sina": 0.24142332599051308, "sagan": 0.21305533739829882, "socrates": 0.032673094202851166}, {"ibn_sina": 0.15379488073095388, "hypatia": 0.017255307395960155, "sagan": 0.09328131047017123, "socrates": 0.03028400760256135, "darwin": 0.4180578296009781, "einstein": 0.28732666419937536}, {"hypatia": 0.02871838096975513, "darwin": 0.5090454827706226, "sagan": 0.0884588409366867, "ibn_sina": 0.11358340009344377, "einstein": 0.24045605988474456, "socrates": 0.019737835344747382}, {"ibn_sina": 0.06342337535655429, "einstein": 0.06342337535655429, "darwin": 0.7726548872486522, "sagan": 0.06342337535655429, "hypatia": 0.029959081180374974, "socrates": 0.00711590550130999}, {"einstein": 0.1857433646787811, "sagan": 0.06030204163840803, "socrates": 0.04696327724879664, "ibn_sina": 0.127659423150288, "darwin": 0.5721298414583078, "hypatia": 0.007202051825418391}, {"socrates": 0.08697393197655402, "ibn_sina": 0.041083576426096335, "hypatia": 0.03625612894312751, "einstein": 0.5005004216950182, "sagan": 0.12654632432147356, "darwin": 0.2086396166377304}], "prompt_tokens": 296, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.3512688666047678, "einstein": 0.11404030424940888, "hypatia": 0.04753907155416017, "ibn_sina": 0.24142332599051308, "sagan": 0.21305533739829882, "socrates": 0.032673094202851166}, "confidence": 0.13814500572755695, "choice": "darwin"}}, "usage": {"input_tokens": 296, "output_tokens": 1}, "latency_ms": 174.53824099720805, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 296, "compiled_prompt_sha256": "36256d9ffe4d50c3d22a8cefe2d22c38863f669314f6c78f6e3165edd736e631", "compiled_ids_sha256": "bedc1945b96a6e27ccb028c42f119ba2eb225cdab520d6ff987d956978f024cd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.0, 17.875, 17.0, 18.625, 18.5, 16.625], "runtime_probabilities": [0.35126885771751404, 0.1140403002500534, 0.04753907024860382, 0.2414233237504959, 0.2130553275346756, 0.03267309069633484], "decide_seconds": 0.1706415120024758, "probabilities_f64_minus_runtime_max_abs": 9.86362322374923e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.15379488073095388, "hypatia": 0.017255307395960155, "sagan": 0.09328131047017123, "socrates": 0.03028400760256135, "darwin": 0.4180578296009781, "einstein": 0.28732666419937536}, "confidence": 0.21412628777795517, "choice": "darwin"}}, "usage": {"input_tokens": 296, "output_tokens": 1}, "latency_ms": 173.5322989989072, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 296, "compiled_prompt_sha256": "e42bd9b0cedfb08ec693aa27f8b2df6de90ca7d55e8463294125dc41439e0a37", "compiled_ids_sha256": "56ee90834bdc5b1bb8decbdf7a4904d5212754d2d6e8373e652dfb8e56f1e39e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 15.9375, 17.625, 16.5, 19.125, 18.75], "runtime_probabilities": [0.15379489958286285, 0.017255308106541634, 0.09328131377696991, 0.030284009873867035, 0.4180578291416168, 0.2873266637325287], "decide_seconds": 0.16995270699771936, "probabilities_f64_minus_runtime_max_abs": 1.885190897832345e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.02871838096975513, "darwin": 0.5090454827706226, "sagan": 0.0884588409366867, "ibn_sina": 0.11358340009344377, "einstein": 0.24045605988474456, "socrates": 0.019737835344747382}, "confidence": 0.25913530305838217, "choice": "darwin"}}, "usage": {"input_tokens": 296, "output_tokens": 1}, "latency_ms": 173.5603990018717, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 296, "compiled_prompt_sha256": "0d3acb839327b5ed8b9a89af5a322a76c70bad66bcaf5f2e160d4bb6c9d57850", "compiled_ids_sha256": "4f983f973454a9880cc52fae24d50b5d2c6b77c1652e37e2e9f1141f3d4860ea", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 19.375, 17.625, 17.875, 18.625, 16.125], "runtime_probabilities": [0.028718382120132446, 0.5090454816818237, 0.0884588286280632, 0.11358340084552765, 0.24045605957508087, 0.019737834110856056], "decide_seconds": 0.17001193600299302, "probabilities_f64_minus_runtime_max_abs": 1.2308623495438376e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.06342337535655429, "einstein": 0.06342337535655429, "darwin": 0.7726548872486522, "sagan": 0.06342337535655429, "hypatia": 0.029959081180374974, "socrates": 0.00711590550130999}, "confidence": 0.5176133069717649, "choice": "darwin"}}, "usage": {"input_tokens": 296, "output_tokens": 1}, "latency_ms": 173.28064999674098, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 296, "compiled_prompt_sha256": "9b6d91cb0b3666a5057a4df12fc8247dda9ac305984b9d9ef3535630cdbab79e", "compiled_ids_sha256": "269ac0cb3bb59fb13b4a76c3d78878c17e480f4302a5ee573041686c723ae9ed", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 17.375, 19.875, 17.375, 16.625, 15.1875], "runtime_probabilities": [0.06342338025569916, 0.06342338025569916, 0.7726549506187439, 0.06342338025569916, 0.02995908260345459, 0.00711590563878417], "decide_seconds": 0.1697528879994934, "probabilities_f64_minus_runtime_max_abs": 6.337009172963093e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.1857433646787811, "sagan": 0.06030204163840803, "socrates": 0.04696327724879664, "ibn_sina": 0.127659423150288, "darwin": 0.5721298414583078, "hypatia": 0.007202051825418391}, "confidence": 0.3060248751609296, "choice": "darwin"}}, "usage": {"input_tokens": 296, "output_tokens": 1}, "latency_ms": 173.1414420028159, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 296, "compiled_prompt_sha256": "c1bb3d6c6a862369b6084f743e6e387dfa5e1afab61d21235ee28083ce7263db", "compiled_ids_sha256": "6d15ea8d0e43ad3952bd2bb8485d0d182cf52eec95bb5533f34210557bdab562", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.25, 17.125, 16.875, 17.875, 19.375, 15.0], "runtime_probabilities": [0.18574336171150208, 0.060302041471004486, 0.04696327820420265, 0.12765942513942719, 0.5721298456192017, 0.007202052045613527], "decide_seconds": 0.16957342000023345, "probabilities_f64_minus_runtime_max_abs": 4.160893851690162e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.08697393197655402, "ibn_sina": 0.041083576426096335, "hypatia": 0.03625612894312751, "einstein": 0.5005004216950182, "sagan": 0.12654632432147356, "darwin": 0.2086396166377304}, "confidence": 0.21931856217155554, "choice": "einstein"}}, "usage": {"input_tokens": 296, "output_tokens": 1}, "latency_ms": 174.11598399849026, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 296, "compiled_prompt_sha256": "4d89b0a4d92b78766d8c6df2302113378d8eabe3158ff0cc3f5de3002ee99f8f", "compiled_ids_sha256": "114a85d8f043627a00e00c3c88c279c6425def97aaa9280ae263fcd587303ed9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 16.75, 16.625, 19.25, 17.875, 18.375], "runtime_probabilities": [0.08697392791509628, 0.041083578020334244, 0.036256130784749985, 0.5005004405975342, 0.12654632329940796, 0.20863962173461914], "decide_seconds": 0.17048716300269007, "probabilities_f64_minus_runtime_max_abs": 1.8902515996899183e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:35Z"} {"id": "c-darwin-01-dessert-02", "uid": "c003", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "prompt_sha256_by_order": ["1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2"], "request_sha256": "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "compiled_prompt_sha256_by_order": ["604afd02ea457c03d0d85cec823cbc75b5a82d677e33fef5295394b9bc38e733", "eecb982988a136fdc7afcc5fa85ebd799ece428ecdb90ce6d98a62ec58e25193", "d4814fd3b0c2bfb52fe0d3c351348f1f63c210a78bcb2c3fd48de52aad02360a", "0c93ef4823810b24557e1073fda065276dfb6e7dd48c59c111f2603bc5b15d83", "0e793654ef3c48245547a933abf15a5952a6c5a6caf5cdd53b831304b7eaee5b", "349c3e96f2945599c0a60343dcd53d3a1781af22f0db26192c8b8b3a0b5b1f4d"], "compiled_tokens_by_order": [399, 399, 399, 399, 399, 399], "choice": "darwin", "probs": {"darwin": 0.948195860864749, "einstein": 0.004391017975137249, "hypatia": 0.006388893455258896, "ibn_sina": 0.00723956473563539, "sagan": 0.03244547814970201, "socrates": 0.0013391848195173878}, "confidence": 0.8535935462339951, "seconds": 0.21102485533386547, "seconds_all_orders": 1.2661491320031928, "seconds_by_order": [0.21103659700020216, 0.21093054799712263, 0.21089057800054434, 0.2105911200014816, 0.2108056690012745, 0.21189462000256754], "argmax_stable": true, "spread": {"darwin": 0.09003996598925335, "einstein": 0.07542991424982683, "hypatia": 0.011350514933662707, "ibn_sina": 0.009670680754382117, "sagan": 0.06011385745477546, "socrates": 0.007385785124999723}, "spread_2dp": {"darwin": 0.08999999999999997, "einstein": 0.08, "hypatia": 0.01, "ibn_sina": 0.0, "sagan": 0.06, "socrates": 0.01}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.948195860864749, "einstein": 0.004391017975137249, "hypatia": 0.006388893455258896, "ibn_sina": 0.00723956473563539, "sagan": 0.03244547814970201, "socrates": 0.0013391848195173878}, {"ibn_sina": 0.01477346223104963, "hypatia": 0.0027328157249772475, "sagan": 0.04015843392608824, "socrates": 0.003976223417575511, "darwin": 0.9140017432780929, "einstein": 0.024357321422216496}, {"hypatia": 0.0071845289096249074, "darwin": 0.9409875901567681, "sagan": 0.02212990699896533, "ibn_sina": 0.011845285633258583, "einstein": 0.015209647821030169, "socrates": 0.0026430404803528826}, {"ibn_sina": 0.010303584817484395, "einstein": 0.020491125197550433, "darwin": 0.8713048147985948, "sagan": 0.08104397229528304, "hypatia": 0.014083330658639955, "socrates": 0.002773172232447379}, {"einstein": 0.06446929238152077, "sagan": 0.020930114840507576, "socrates": 0.008724969944517111, "ibn_sina": 0.012694756362074242, "darwin": 0.8899711294039537, "hypatia": 0.003209737067426585}, {"socrates": 0.00841306436031634, "ibn_sina": 0.005102781476667514, "hypatia": 0.010802588470875977, "einstein": 0.07982093222496407, "sagan": 0.03770473859168045, "darwin": 0.8581558948754957}], "prompt_tokens": 399, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.948195860864749, "einstein": 0.004391017975137249, "hypatia": 0.006388893455258896, "ibn_sina": 0.00723956473563539, "sagan": 0.03244547814970201, "socrates": 0.0013391848195173878}, "confidence": 0.8535935462339951, "choice": "darwin"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 211.00833699892974, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 399, "compiled_prompt_sha256": "604afd02ea457c03d0d85cec823cbc75b5a82d677e33fef5295394b9bc38e733", "compiled_ids_sha256": "77f997cd82656949ba572af316c23fd06bc92bfc51d6b9880262378e26c5a2fd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.875, 15.5, 15.875, 16.0, 17.5, 14.3125], "runtime_probabilities": [0.9481959342956543, 0.004391018766909838, 0.0063888938166201115, 0.007239564787596464, 0.032445479184389114, 0.0013391849352046847], "decide_seconds": 0.20619998399706674, "probabilities_f64_minus_runtime_max_abs": 7.343090524525309e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.01477346223104963, "hypatia": 0.0027328157249772475, "sagan": 0.04015843392608824, "socrates": 0.003976223417575511, "darwin": 0.9140017432780929, "einstein": 0.024357321422216496}, "confidence": 0.7755508251848344, "choice": "darwin"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 210.90505800020765, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 399, "compiled_prompt_sha256": "eecb982988a136fdc7afcc5fa85ebd799ece428ecdb90ce6d98a62ec58e25193", "compiled_ids_sha256": "e07465fed92b998815c200a700fdd85da9cfc4c20951a7dbaa1575588e4be7dc", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 14.4375, 17.125, 14.8125, 20.25, 16.625], "runtime_probabilities": [0.014773461036384106, 0.0027328156866133213, 0.040158431977033615, 0.003976223058998585, 0.9140017032623291, 0.024357320740818977], "decide_seconds": 0.2061168260006525, "probabilities_f64_minus_runtime_max_abs": 4.0015763813627814e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0071845289096249074, "darwin": 0.9409875901567681, "sagan": 0.02212990699896533, "ibn_sina": 0.011845285633258583, "einstein": 0.015209647821030169, "socrates": 0.0026430404803528826}, "confidence": 0.8275840886400585, "choice": "darwin"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 210.86563800054137, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 399, "compiled_prompt_sha256": "d4814fd3b0c2bfb52fe0d3c351348f1f63c210a78bcb2c3fd48de52aad02360a", "compiled_ids_sha256": "7f9a490fafd5512098bba5444a69e59e3122a02dc5193c4ac061e78967177a72", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.875, 20.75, 17.0, 16.375, 16.625, 14.875], "runtime_probabilities": [0.00718452874571085, 0.9409875869750977, 0.022129908204078674, 0.011845286004245281, 0.015209647826850414, 0.0026430406142026186], "decide_seconds": 0.2062141749993316, "probabilities_f64_minus_runtime_max_abs": 3.181670482632626e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.010303584817484395, "einstein": 0.020491125197550433, "darwin": 0.8713048147985948, "sagan": 0.08104397229528304, "hypatia": 0.014083330658639955, "socrates": 0.002773172232447379}, "confidence": 0.7059615403532482, "choice": "darwin"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 210.56555000177468, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 399, "compiled_prompt_sha256": "0c93ef4823810b24557e1073fda065276dfb6e7dd48c59c111f2603bc5b15d83", "compiled_ids_sha256": "f07dedf64fdfc9daf35a77d0235b5d2fa815c708313d843e26184b6a70649a04", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 16.625, 20.375, 18.0, 16.25, 14.625], "runtime_probabilities": [0.010303584858775139, 0.02049112692475319, 0.8713048696517944, 0.08104397356510162, 0.01408333145081997, 0.00277317245490849], "decide_seconds": 0.20586363700203947, "probabilities_f64_minus_runtime_max_abs": 5.485319964915192e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.06446929238152077, "sagan": 0.020930114840507576, "socrates": 0.008724969944517111, "ibn_sina": 0.012694756362074242, "darwin": 0.8899711294039537, "hypatia": 0.003209737067426585}, "confidence": 0.7339784951440518, "choice": "darwin"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 210.77968900135602, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 399, "compiled_prompt_sha256": "0e793654ef3c48245547a933abf15a5952a6c5a6caf5cdd53b831304b7eaee5b", "compiled_ids_sha256": "ef5fbcd2e719aeca70027a5142e9f5a5f819a3c42c9935b2f27fe7f883518c5c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 16.5, 15.625, 16.0, 20.25, 14.625], "runtime_probabilities": [0.06446929275989532, 0.02093011513352394, 0.008724970743060112, 0.01269475556910038, 0.889971137046814, 0.003209737129509449], "decide_seconds": 0.20614125599968247, "probabilities_f64_minus_runtime_max_abs": 7.642860233580961e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.00841306436031634, "ibn_sina": 0.005102781476667514, "hypatia": 0.010802588470875977, "einstein": 0.07982093222496407, "sagan": 0.03770473859168045, "darwin": 0.8581558948754957}, "confidence": 0.6803728209042266, "choice": "darwin"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 211.8667210015701, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 399, "compiled_prompt_sha256": "349c3e96f2945599c0a60343dcd53d3a1781af22f0db26192c8b8b3a0b5b1f4d", "compiled_ids_sha256": "d3752dbfec8ff57afd0889338a3cbf8b0c19159b4effc6f1fc6706057e3787c1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.875, 15.375, 16.125, 18.125, 17.375, 20.5], "runtime_probabilities": [0.00841306522488594, 0.00510278157889843, 0.010802589356899261, 0.07982093095779419, 0.03770473971962929, 0.8581559062004089], "decide_seconds": 0.2066597709999769, "probabilities_f64_minus_runtime_max_abs": 1.1324913229415756e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:36Z"} {"id": "c-darwin-01-dessert-03", "uid": "c004", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "prompt_sha256_by_order": ["7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c"], "request_sha256": "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "compiled_prompt_sha256_by_order": ["f23455f0797804d5728aeff3fdc9d48c4b31d4a7b4b5aa74d195e7140ad387f6", "9fa8721a8c13ea09d37498a6c5184e3a2a581f98cac48c4886410653bc6b6ad0", "7f0f69b8f7631a215fecc71cfdd27c835cfcca1a7c03c3ef269839607f7d3091", "fed544507f2d46f84c7f22c5fec1bbfee70fe8aa08be2e63ec0d02302847bcba", "56018aedd0d7a17a9d82f95e083c24bad6beb6d0a716cf2c5009462115797d77", "cdd784ffeb6b17ba13fea7f684921b658c3adf1c0f1b9ddb2b01afe5a57952ff"], "compiled_tokens_by_order": [409, 409, 409, 409, 409, 409], "choice": "darwin", "probs": {"darwin": 0.37073983528176097, "einstein": 0.03907569115289092, "hypatia": 0.1363875634234468, "ibn_sina": 0.1545473565108324, "sagan": 0.28873247403319907, "socrates": 0.010517079597869863}, "confidence": 0.1843538376013686, "seconds": 0.212126606833408, "seconds_all_orders": 1.272759641000448, "seconds_by_order": [0.2124667260031856, 0.2124681759996747, 0.21182637099991553, 0.21227448699937668, 0.2119296300006681, 0.2117942509976274], "argmax_stable": false, "spread": {"darwin": 0.5049941803289776, "einstein": 0.34706138572144885, "hypatia": 0.1990296062798799, "ibn_sina": 0.1335553827468644, "sagan": 0.3840956088835257, "socrates": 0.017746112597219463}, "spread_2dp": {"darwin": 0.5, "einstein": 0.33999999999999997, "hypatia": 0.2, "ibn_sina": 0.13, "sagan": 0.39, "socrates": 0.019999999999999997}, "choices_by_order": ["darwin", "sagan", "darwin", "darwin", "darwin", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.37073983528176097, "einstein": 0.03907569115289092, "hypatia": 0.1363875634234468, "ibn_sina": 0.1545473565108324, "sagan": 0.28873247403319907, "socrates": 0.010517079597869863}, {"ibn_sina": 0.038783349720221495, "hypatia": 0.026655380458309084, "sagan": 0.416960304242312, "socrates": 0.014267597021831696, "darwin": 0.3679661769945709, "einstein": 0.13536719156275479}, {"hypatia": 0.031285239693582836, "darwin": 0.6283808370162346, "sagan": 0.1402107167976368, "ibn_sina": 0.06623085295104637, "einstein": 0.12373552328308038, "socrates": 0.010156830258419049}, {"ibn_sina": 0.06803928023750273, "einstein": 0.025030252391479664, "darwin": 0.44367183950620653, "sagan": 0.26910057351161704, "hypatia": 0.18494993909103627, "socrates": 0.009208115262157696}, {"einstein": 0.08933550419163125, "sagan": 0.03286469535878635, "socrates": 0.022587552770825275, "ibn_sina": 0.04781784959107677, "darwin": 0.6601050520982176, "hypatia": 0.1472893459894629}, {"socrates": 0.02695422785937716, "ibn_sina": 0.02099197376396801, "hypatia": 0.225684986738189, "einstein": 0.3720916381129285, "sagan": 0.19916630175629726, "darwin": 0.15511087176924002}], "prompt_tokens": 409, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.37073983528176097, "einstein": 0.03907569115289092, "hypatia": 0.1363875634234468, "ibn_sina": 0.1545473565108324, "sagan": 0.28873247403319907, "socrates": 0.010517079597869863}, "confidence": 0.1843538376013686, "choice": "darwin"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 212.43653599958634, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "f23455f0797804d5728aeff3fdc9d48c4b31d4a7b4b5aa74d195e7140ad387f6", "compiled_ids_sha256": "2fc9c2e9c165a1094386aba821c96f2b186d61c9c48cddddeb2a7985bc7f809a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.375, 17.125, 18.375, 18.5, 19.125, 15.8125], "runtime_probabilities": [0.3707398772239685, 0.03907569497823715, 0.13638758659362793, 0.15454736351966858, 0.28873249888420105, 0.010517081245779991], "decide_seconds": 0.20738710500154411, "probabilities_f64_minus_runtime_max_abs": 4.19422075403908e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.038783349720221495, "hypatia": 0.026655380458309084, "sagan": 0.416960304242312, "socrates": 0.014267597021831696, "darwin": 0.3679661769945709, "einstein": 0.13536719156275479}, "confidence": 0.2819270488580973, "choice": "sagan"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 212.44224600013695, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "9fa8721a8c13ea09d37498a6c5184e3a2a581f98cac48c4886410653bc6b6ad0", "compiled_ids_sha256": "d6914ad33a20e73683ad4646cfbf5dd65a305eced249bf7a118714060ab9afc5", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 16.625, 19.375, 16.0, 19.25, 18.25], "runtime_probabilities": [0.038783349096775055, 0.02665538154542446, 0.41696032881736755, 0.01426759734749794, 0.3679662048816681, 0.13536719977855682], "decide_seconds": 0.20762783300233423, "probabilities_f64_minus_runtime_max_abs": 2.788709718704041e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.031285239693582836, "darwin": 0.6283808370162346, "sagan": 0.1402107167976368, "ibn_sina": 0.06623085295104637, "einstein": 0.12373552328308038, "socrates": 0.010156830258419049}, "confidence": 0.3521634731386948, "choice": "darwin"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 211.8009309997433, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "7f0f69b8f7631a215fecc71cfdd27c835cfcca1a7c03c3ef269839607f7d3091", "compiled_ids_sha256": "79ca49438526ba73e13ce1ce1676b217ae575e085426773c613a7e87efcee660", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 19.875, 18.375, 17.625, 18.25, 15.75], "runtime_probabilities": [0.03128524124622345, 0.6283808350563049, 0.14021073281764984, 0.06623085588216782, 0.12373551726341248, 0.01015683077275753], "decide_seconds": 0.20697297899823752, "probabilities_f64_minus_runtime_max_abs": 1.6020013038353653e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.06803928023750273, "einstein": 0.025030252391479664, "darwin": 0.44367183950620653, "sagan": 0.26910057351161704, "hypatia": 0.18494993909103627, "socrates": 0.009208115262157696}, "confidence": 0.24974938361471166, "choice": "darwin"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 212.24948700182722, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "fed544507f2d46f84c7f22c5fec1bbfee70fe8aa08be2e63ec0d02302847bcba", "compiled_ids_sha256": "e3ce0791ee677404ad2ba5e76cad11df8fcee12777f77958466d2dbbb4a33b6f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 16.375, 19.25, 18.75, 18.375, 15.375], "runtime_probabilities": [0.06803928315639496, 0.02503025345504284, 0.4436718225479126, 0.2691005766391754, 0.18494993448257446, 0.009208114817738533], "decide_seconds": 0.20753881400014507, "probabilities_f64_minus_runtime_max_abs": 1.695829393577597e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.08933550419163125, "sagan": 0.03286469535878635, "socrates": 0.022587552770825275, "ibn_sina": 0.04781784959107677, "darwin": 0.6601050520982176, "hypatia": 0.1472893459894629}, "confidence": 0.3775334998036901, "choice": "darwin"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 211.9044799983385, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "56018aedd0d7a17a9d82f95e083c24bad6beb6d0a716cf2c5009462115797d77", "compiled_ids_sha256": "24863b7310897a5b13d6d6ac85f5b3f83351e7a1c22332e9ba4d2e875f7cb554", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 16.75, 16.375, 17.125, 19.75, 18.25], "runtime_probabilities": [0.08933550119400024, 0.03286469355225563, 0.022587550804018974, 0.04781784862279892, 0.6601050496101379, 0.14728933572769165], "decide_seconds": 0.20721457699983148, "probabilities_f64_minus_runtime_max_abs": 1.026177123830685e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.02695422785937716, "ibn_sina": 0.02099197376396801, "hypatia": 0.225684986738189, "einstein": 0.3720916381129285, "sagan": 0.19916630175629726, "darwin": 0.15511087176924002}, "confidence": 0.16687180442669458, "choice": "einstein"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 211.76862100037397, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "cdd784ffeb6b17ba13fea7f684921b658c3adf1c0f1b9ddb2b01afe5a57952ff", "compiled_ids_sha256": "cd690b2a8827e88dec9d609a4e2afab20cdeb3985dc7b8c97f0fc407b0ab1094", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 16.375, 18.75, 19.25, 18.625, 18.375], "runtime_probabilities": [0.026954228058457375, 0.02099197544157505, 0.2256849855184555, 0.3720916509628296, 0.19916631281375885, 0.15511086583137512], "decide_seconds": 0.20688053900084924, "probabilities_f64_minus_runtime_max_abs": 1.28499010876304e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:37Z"} {"id": "c-darwin-01-dessert-03", "uid": "c005", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "prompt_sha256_by_order": ["1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06"], "request_sha256": "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "compiled_prompt_sha256_by_order": ["24e4de0b01d855575257e46ab574c911b12ec716af49ef43360d71b0ceb39554", "7ae82cf86dec5f8bd087468bc1c0b3049d4f35086b412a067c128b307da5d00b", "bf6d464337d82fb63d386fdea062a1714e3f546ce4e7485345e36242709deb27", "980e3454da994c6c2b3c6b12697cde2623b178247de8e4b821e04ce52bffdea4", "cb00dc996172d719bcdc935c287e628ad41efa192010278591ebc2679d4083e0", "0d4e9e0115d996b05648a1b53ae79f2fe148aa38b656e90f7f7f749a2c6f3051"], "compiled_tokens_by_order": [411, 411, 411, 411, 411, 411], "choice": "darwin", "probs": {"darwin": 0.9969224490920827, "einstein": 0.00040339985321802014, "hypatia": 0.0005869433230904619, "ibn_sina": 0.0011672761791511467, "sagan": 0.0006247979151286316, "socrates": 0.0002951336373290848}, "confidence": 0.9857769925598754, "seconds": 0.2127236655002586, "seconds_all_orders": 1.2763419930015516, "seconds_by_order": [0.21322353000141447, 0.21285035300024902, 0.21232563699959428, 0.21268753399999696, 0.21261321500060149, 0.2126417239996954], "argmax_stable": true, "spread": {"darwin": 0.008802302105621718, "einstein": 0.004179953091142766, "hypatia": 0.001099177996772469, "ibn_sina": 0.006377114242644238, "sagan": 0.0007549301393617223, "socrates": 0.0037456676086950694}, "spread_2dp": {"darwin": 0.010000000000000009, "einstein": 0.0, "hypatia": 0.0, "ibn_sina": 0.01, "sagan": 0.0, "socrates": 0.0}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.9969224490920827, "einstein": 0.00040339985321802014, "hypatia": 0.0005869433230904619, "ibn_sina": 0.0011672761791511467, "sagan": 0.0006247979151286316, "socrates": 0.0002951336373290848}, {"ibn_sina": 0.007544390421795384, "hypatia": 0.001156969742999627, "sagan": 0.0008464571461969066, "socrates": 0.0008464571461969066, "darwin": 0.988120146986461, "einstein": 0.001485578556350205}, {"hypatia": 0.0013206608441686224, "darwin": 0.9953876416057535, "sagan": 0.0004027788007828312, "ibn_sina": 0.0013206608441686224, "einstein": 0.0011654791043435662, "socrates": 0.0004027788007828312}, {"ibn_sina": 0.002971120942466183, "einstein": 0.0007996663532975847, "darwin": 0.9937039281807963, "sagan": 0.0009061406912087485, "hypatia": 0.0011635076786070313, "socrates": 0.00045563615362417413}, {"einstein": 0.004583352944360786, "sagan": 0.0007482118001595356, "socrates": 0.0010886417455481536, "ibn_sina": 0.0021650226303230845, "darwin": 0.9897286495597455, "hypatia": 0.001686121319862931}, {"socrates": 0.004040801246024154, "ibn_sina": 0.0013964635436274641, "hypatia": 0.001087566901307738, "einstein": 0.00356599458357629, "sagan": 0.0011577089401445535, "darwin": 0.9887514647853198}], "prompt_tokens": 411, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.9969224490920827, "einstein": 0.00040339985321802014, "hypatia": 0.0005869433230904619, "ibn_sina": 0.0011672761791511467, "sagan": 0.0006247979151286316, "socrates": 0.0002951336373290848}, "confidence": 0.9857769925598754, "choice": "darwin"}}, "usage": {"input_tokens": 411, "output_tokens": 1}, "latency_ms": 213.19505000064964, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 411, "compiled_prompt_sha256": "24e4de0b01d855575257e46ab574c911b12ec716af49ef43360d71b0ceb39554", "compiled_ids_sha256": "7edad3a35e2849ccf2f8adce479f5cd518ba5078d299c95eded9c3ee2725b892", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.75, 12.9375, 13.3125, 14.0, 13.375, 12.625], "runtime_probabilities": [0.9969224333763123, 0.0004033998411614448, 0.0005869432934559882, 0.0011672761756926775, 0.0006247978890314698, 0.0002951336500700563], "decide_seconds": 0.20817081999848597, "probabilities_f64_minus_runtime_max_abs": 1.571577046277639e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.007544390421795384, "hypatia": 0.001156969742999627, "sagan": 0.0008464571461969066, "socrates": 0.0008464571461969066, "darwin": 0.988120146986461, "einstein": 0.001485578556350205}, "confidence": 0.9563825945468066, "choice": "darwin"}}, "usage": {"input_tokens": 411, "output_tokens": 1}, "latency_ms": 212.8242130020226, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 411, "compiled_prompt_sha256": "7ae82cf86dec5f8bd087468bc1c0b3049d4f35086b412a067c128b307da5d00b", "compiled_ids_sha256": "338b01856f416d4a61dc12810e036acb1a8a1cb1ba05c9d1f761350fa5bc444b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.375, 13.5, 13.1875, 13.1875, 20.25, 13.75], "runtime_probabilities": [0.007544389460235834, 0.0011569696944206953, 0.0008464570855721831, 0.0008464570855721831, 0.9881200790405273, 0.0014855784829705954], "decide_seconds": 0.20792350099873147, "probabilities_f64_minus_runtime_max_abs": 6.794593365722079e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0013206608441686224, "darwin": 0.9953876416057535, "sagan": 0.0004027788007828312, "ibn_sina": 0.0013206608441686224, "einstein": 0.0011654791043435662, "socrates": 0.0004027788007828312}, "confidence": 0.9797505258572765, "choice": "darwin"}}, "usage": {"input_tokens": 411, "output_tokens": 1}, "latency_ms": 212.29880700047943, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 411, "compiled_prompt_sha256": "bf6d464337d82fb63d386fdea062a1714e3f546ce4e7485345e36242709deb27", "compiled_ids_sha256": "d14b25d3db65b01907cc3f123ee7db381c4af25471f371b300083463f3006978", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.625, 21.25, 13.4375, 14.625, 14.5, 13.4375], "runtime_probabilities": [0.0013206609291955829, 0.9953877329826355, 0.000402778823627159, 0.0013206609291955829, 0.001165479188784957, 0.000402778823627159], "decide_seconds": 0.20733067600303912, "probabilities_f64_minus_runtime_max_abs": 9.137688194993387e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.002971120942466183, "einstein": 0.0007996663532975847, "darwin": 0.9937039281807963, "sagan": 0.0009061406912087485, "hypatia": 0.0011635076786070313, "socrates": 0.00045563615362417413}, "confidence": 0.9737785097164768, "choice": "darwin"}}, "usage": {"input_tokens": 411, "output_tokens": 1}, "latency_ms": 212.66122400265886, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 411, "compiled_prompt_sha256": "980e3454da994c6c2b3c6b12697cde2623b178247de8e4b821e04ce52bffdea4", "compiled_ids_sha256": "c2279aafb7c24ad5b099f1127ff01aa5cafc0d9df0a980f9fc1cefa9cd26d047", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.6875, 13.375, 20.5, 13.5, 13.75, 12.8125], "runtime_probabilities": [0.0029711208771914244, 0.0007996663334779441, 0.9937039017677307, 0.0009061406599357724, 0.001163507578894496, 0.0004556361527647823], "decide_seconds": 0.2077218129998073, "probabilities_f64_minus_runtime_max_abs": 2.6413065623565046e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.004583352944360786, "sagan": 0.0007482118001595356, "socrates": 0.0010886417455481536, "ibn_sina": 0.0021650226303230845, "darwin": 0.9897286495597455, "hypatia": 0.001686121319862931}, "confidence": 0.9599477947148229, "choice": "darwin"}}, "usage": {"input_tokens": 411, "output_tokens": 1}, "latency_ms": 212.58670500174048, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 411, "compiled_prompt_sha256": "cb00dc996172d719bcdc935c287e628ad41efa192010278591ebc2679d4083e0", "compiled_ids_sha256": "fb3d3717bd9882bff0d11c1fec042a2cfbea9ff0407970b3aea1ec745500632c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0, 13.1875, 13.5625, 14.25, 20.375, 14.0], "runtime_probabilities": [0.00458335317671299, 0.0007482117507606745, 0.001088641700334847, 0.002165022538974881, 0.989728569984436, 0.0016861213371157646], "decide_seconds": 0.20752606399764773, "probabilities_f64_minus_runtime_max_abs": 7.957530945290614e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.004040801246024154, "ibn_sina": 0.0013964635436274641, "hypatia": 0.001087566901307738, "einstein": 0.00356599458357629, "sagan": 0.0011577089401445535, "darwin": 0.9887514647853198}, "confidence": 0.9564767017889645, "choice": "darwin"}}, "usage": {"input_tokens": 411, "output_tokens": 1}, "latency_ms": 212.61570500064408, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 411, "compiled_prompt_sha256": "0d4e9e0115d996b05648a1b53ae79f2fe148aa38b656e90f7f7f749a2c6f3051", "compiled_ids_sha256": "4ecfe7f6c8cad67259c06bd9e8a387fa43c38e9c23586ef1d7e4501f297ed7ef", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0, 13.9375, 13.6875, 14.875, 13.75, 20.5], "runtime_probabilities": [0.004040801431983709, 0.0013964636018499732, 0.0010875668376684189, 0.003565994556993246, 0.0011577089317142963, 0.9887514710426331], "decide_seconds": 0.20766433400058304, "probabilities_f64_minus_runtime_max_abs": 6.257313223123617e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:38Z"} {"id": "c-darwin-01-main-01", "uid": "c006", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "prompt_sha256_by_order": ["56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545"], "request_sha256": "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "compiled_prompt_sha256_by_order": ["377f0fd9228b401ffd1752b4418d7a4f949f5c9987b66900bf4451b634fa3b16", "8e484681f634b2f4c1cd7123121993b1b27afe0c195f941c7734f878d0bbd2fd", "7878a73c4f5ae760d37908f3a6ceb98ac84c9a1a487cce85de596986d9a825d8", "9c31b0fac9c51e308e2ab23a9c928df2a8a5f926e559d266b4f0aae4a6dd586b", "727c8ae7ff231dce223ce04956fa1362ac85079812affa37eae8833c4e87dec2", "7ab880c611545781cc4584e295dca04b32e54cbfd963b82802eff6a3c2ca1ddd"], "compiled_tokens_by_order": [314, 314, 314, 314, 314, 314], "choice": "darwin", "probs": {"darwin": 0.7293555439716106, "einstein": 0.015137291389642876, "hypatia": 0.02828018344517389, "ibn_sina": 0.05986914882324262, "sagan": 0.16274121933153063, "socrates": 0.0046166130387993625}, "confidence": 0.507010917874219, "seconds": 0.17644613616539573, "seconds_all_orders": 1.0586768169923744, "seconds_by_order": [0.17687295300129335, 0.1761461789974419, 0.176323086998309, 0.17677003299831995, 0.17593494999891846, 0.1766296149980917], "argmax_stable": true, "spread": {"darwin": 0.22220485098818166, "einstein": 0.0817845820380004, "hypatia": 0.02758902915475788, "ibn_sina": 0.055298773508335244, "sagan": 0.12221540756303337, "socrates": 0.025434771379293127}, "spread_2dp": {"darwin": 0.22999999999999998, "einstein": 0.08, "hypatia": 0.019999999999999997, "ibn_sina": 0.05, "sagan": 0.12, "socrates": 0.03}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.7293555439716106, "einstein": 0.015137291389642876, "hypatia": 0.02828018344517389, "ibn_sina": 0.05986914882324262, "sagan": 0.16274121933153063, "socrates": 0.0046166130387993625}, {"ibn_sina": 0.03870822708407906, "hypatia": 0.005936098881104906, "sagan": 0.04970234741087834, "socrates": 0.010418177985459983, "darwin": 0.8809951876900495, "einstein": 0.014239960948428292}, {"hypatia": 0.033525128035862786, "darwin": 0.6733701970189161, "sagan": 0.11701419453798022, "ibn_sina": 0.08042260137231494, "einstein": 0.09113074633664871, "socrates": 0.004537132698277195}, {"ibn_sina": 0.02704402310813163, "einstein": 0.009346164298648298, "darwin": 0.8955750480070978, "sagan": 0.050524886844470526, "hypatia": 0.014475622441759118, "socrates": 0.0030342552998926906}, {"einstein": 0.06681576787511102, "sagan": 0.04052581176849727, "socrates": 0.019143038002116298, "ibn_sina": 0.05203617234265368, "darwin": 0.8139826886183053, "hypatia": 0.007496521393316533}, {"socrates": 0.028469026679185817, "ibn_sina": 0.02512382786397969, "hypatia": 0.02512382786397969, "einstein": 0.0876907756495251, "sagan": 0.0993666667754895, "darwin": 0.7342258751678402}], "prompt_tokens": 314, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.7293555439716106, "einstein": 0.015137291389642876, "hypatia": 0.02828018344517389, "ibn_sina": 0.05986914882324262, "sagan": 0.16274121933153063, "socrates": 0.0046166130387993625}, "confidence": 0.507010917874219, "choice": "darwin"}}, "usage": {"input_tokens": 314, "output_tokens": 1}, "latency_ms": 176.84392299997853, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 314, "compiled_prompt_sha256": "377f0fd9228b401ffd1752b4418d7a4f949f5c9987b66900bf4451b634fa3b16", "compiled_ids_sha256": "01e1272f80d0db6c2b279d6449fa670109af77f7a80b594d0041b4894a23f3be", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.0, 16.125, 16.75, 17.5, 18.5, 14.9375], "runtime_probabilities": [0.72935551404953, 0.015137290582060814, 0.02828018181025982, 0.05986914411187172, 0.16274121403694153, 0.004616612568497658], "decide_seconds": 0.17291388399826246, "probabilities_f64_minus_runtime_max_abs": 2.9922080591049394e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.03870822708407906, "hypatia": 0.005936098881104906, "sagan": 0.04970234741087834, "socrates": 0.010418177985459983, "darwin": 0.8809951876900495, "einstein": 0.014239960948428292}, "confidence": 0.7068738405014314, "choice": "darwin"}}, "usage": {"input_tokens": 314, "output_tokens": 1}, "latency_ms": 176.12055900099222, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 314, "compiled_prompt_sha256": "8e484681f634b2f4c1cd7123121993b1b27afe0c195f941c7734f878d0bbd2fd", "compiled_ids_sha256": "bb9f7932058a5211dd00b00544533586a3bca0139e18ab3f9af64d9bcb455221", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 15.0, 17.125, 15.5625, 20.0, 15.875], "runtime_probabilities": [0.03870822861790657, 0.005936098750680685, 0.04970235005021095, 0.010418177582323551, 0.8809952139854431, 0.0142399612814188], "decide_seconds": 0.17234922799980268, "probabilities_f64_minus_runtime_max_abs": 2.6295393640296538e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.033525128035862786, "darwin": 0.6733701970189161, "sagan": 0.11701419453798022, "ibn_sina": 0.08042260137231494, "einstein": 0.09113074633664871, "socrates": 0.004537132698277195}, "confidence": 0.3991071581725848, "choice": "darwin"}}, "usage": {"input_tokens": 314, "output_tokens": 1}, "latency_ms": 176.29854799815803, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 314, "compiled_prompt_sha256": "7878a73c4f5ae760d37908f3a6ceb98ac84c9a1a487cce85de596986d9a825d8", "compiled_ids_sha256": "a1267f356df856ccb6a9b8ce3f42b133d01c199f5bd9d78edee44e5a39f9f0be", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 19.75, 18.0, 17.625, 17.75, 14.75], "runtime_probabilities": [0.03352512791752815, 0.6733702421188354, 0.11701419204473495, 0.08042261004447937, 0.09113074839115143, 0.0045371330343186855], "decide_seconds": 0.17247473699899274, "probabilities_f64_minus_runtime_max_abs": 4.509991935464086e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.02704402310813163, "einstein": 0.009346164298648298, "darwin": 0.8955750480070978, "sagan": 0.050524886844470526, "hypatia": 0.014475622441759118, "socrates": 0.0030342552998926906}, "confidence": 0.7377920763820456, "choice": "darwin"}}, "usage": {"input_tokens": 314, "output_tokens": 1}, "latency_ms": 176.74502400041092, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 314, "compiled_prompt_sha256": "9c31b0fac9c51e308e2ab23a9c928df2a8a5f926e559d266b4f0aae4a6dd586b", "compiled_ids_sha256": "6db79f55d0ee6ea016effba61e5d8207f47d881faf68f4a0af77584c3c5a6554", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 15.5625, 20.125, 17.25, 16.0, 14.4375], "runtime_probabilities": [0.02704402431845665, 0.009346164762973785, 0.8955750465393066, 0.050524890422821045, 0.014475622214376926, 0.0030342554673552513], "decide_seconds": 0.17281361500135972, "probabilities_f64_minus_runtime_max_abs": 3.578350518673723e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.06681576787511102, "sagan": 0.04052581176849727, "socrates": 0.019143038002116298, "ibn_sina": 0.05203617234265368, "darwin": 0.8139826886183053, "hypatia": 0.007496521393316533}, "confidence": 0.5845096071657248, "choice": "darwin"}}, "usage": {"input_tokens": 314, "output_tokens": 1}, "latency_ms": 175.90895999819622, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 314, "compiled_prompt_sha256": "727c8ae7ff231dce223ce04956fa1362ac85079812affa37eae8833c4e87dec2", "compiled_ids_sha256": "f3fb0b57ba2996e7cb8625328b93d66b8b4602ed8f5d038b5404cc574cb40439", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 16.875, 16.125, 17.125, 19.875, 15.1875], "runtime_probabilities": [0.06681577116250992, 0.04052581265568733, 0.019143039360642433, 0.05203617364168167, 0.8139827251434326, 0.007496521342545748], "decide_seconds": 0.1719638809991011, "probabilities_f64_minus_runtime_max_abs": 3.652512736707081e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.028469026679185817, "ibn_sina": 0.02512382786397969, "hypatia": 0.02512382786397969, "einstein": 0.0876907756495251, "sagan": 0.0993666667754895, "darwin": 0.7342258751678402}, "confidence": 0.46637649255543134, "choice": "darwin"}}, "usage": {"input_tokens": 314, "output_tokens": 1}, "latency_ms": 176.60511500071152, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 314, "compiled_prompt_sha256": "7ab880c611545781cc4584e295dca04b32e54cbfd963b82802eff6a3c2ca1ddd", "compiled_ids_sha256": "8b611d8553e8f365c2aad1be5e33903179b0ffdd2caa1f5940bf817d74748658", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 16.5, 16.5, 17.75, 17.875, 19.875], "runtime_probabilities": [0.02846902795135975, 0.025123829022049904, 0.025123829022049904, 0.08769077807664871, 0.0993666723370552, 0.7342259287834167], "decide_seconds": 0.17277403399930336, "probabilities_f64_minus_runtime_max_abs": 5.361557653404958e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:40Z"} {"id": "c-darwin-01-main-02", "uid": "c007", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "prompt_sha256_by_order": ["7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf"], "request_sha256": "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "compiled_prompt_sha256_by_order": ["98b6551eb87406f35159db42e99ab68b6616d9349e25272937a35b9bb812bc44", "a5cdaad818d34b9a9d13e3f249d193a4388ffe12a104296b34eeedb316bd586d", "e8618ebc3c93376c0d6c137dc53cca0882fcc0f2c62737655841c3f0f9f0e650", "8d3c1b75fc680d29a518eb5fa9bbda1574c0e9b4b3b1b2250a450711dc27cc9f", "d2bea664d466f575873c7433bfa2ddd0d814a958f44532e762cb9be6a106ac51", "612b973b2241207c94ab421257e98cc6d1faba4ecac98ae2e27795cd7dcb2c6d"], "compiled_tokens_by_order": [295, 295, 295, 295, 295, 295], "choice": "socrates", "probs": {"darwin": 0.16495688389698435, "einstein": 0.022324486604027412, "hypatia": 0.24001084985228366, "ibn_sina": 0.1455739390990954, "sagan": 0.07792011459731829, "socrates": 0.34921372595029093}, "confidence": 0.12296119299580588, "seconds": 0.17388236133350196, "seconds_all_orders": 1.0432941680010117, "seconds_by_order": [0.1739296359992295, 0.17350226900089183, 0.17361081800117972, 0.17345845000090776, 0.17450936200111755, 0.17428363299768534], "argmax_stable": true, "spread": {"darwin": 0.24515412244433782, "einstein": 0.02641269444593011, "hypatia": 0.22215945511397436, "ibn_sina": 0.21412103200412344, "sagan": 0.06650271737598443, "socrates": 0.43467694643620064}, "spread_2dp": {"darwin": 0.25, "einstein": 0.030000000000000002, "hypatia": 0.22, "ibn_sina": 0.21, "sagan": 0.07, "socrates": 0.43999999999999995}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.16495688389698435, "einstein": 0.022324486604027412, "hypatia": 0.24001084985228366, "ibn_sina": 0.1455739390990954, "sagan": 0.07792011459731829, "socrates": 0.34921372595029093}, {"ibn_sina": 0.10230179331729657, "hypatia": 0.09028101573136368, "sagan": 0.03763472655640346, "socrates": 0.4584848289861117, "darwin": 0.27808510579318024, "einstein": 0.03321252961564426}, {"hypatia": 0.24750782806259655, "darwin": 0.07091218000038109, "sagan": 0.037956554766399746, "ibn_sina": 0.07091218000038109, "einstein": 0.04873718104995752, "socrates": 0.5239740761202839}, {"ibn_sina": 0.2397677076235133, "einstein": 0.02527133045849411, "darwin": 0.1867312784524278, "sagan": 0.03676956885308838, "hypatia": 0.2397677076235133, "socrates": 0.2716924069889632}, {"einstein": 0.03985063198402705, "sagan": 0.011417397221333864, "socrates": 0.7063693534251638, "ibn_sina": 0.051169224338560165, "darwin": 0.09559669651545752, "hypatia": 0.09559669651545752}, {"socrates": 0.5837156465950228, "ibn_sina": 0.025646675619389863, "hypatia": 0.31244047084533805, "einstein": 0.022633111795703414, "sagan": 0.022633111795703414, "darwin": 0.03293098334884241}], "prompt_tokens": 295, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.16495688389698435, "einstein": 0.022324486604027412, "hypatia": 0.24001084985228366, "ibn_sina": 0.1455739390990954, "sagan": 0.07792011459731829, "socrates": 0.34921372595029093}, "confidence": 0.12296119299580588, "choice": "socrates"}}, "usage": {"input_tokens": 295, "output_tokens": 1}, "latency_ms": 173.9038060004532, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 295, "compiled_prompt_sha256": "98b6551eb87406f35159db42e99ab68b6616d9349e25272937a35b9bb812bc44", "compiled_ids_sha256": "1e5073f8ebc6dc8593ad0d9a35127317ad2b076509d4a0a08bc1f851b7949344", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 16.125, 18.5, 18.0, 17.375, 18.875], "runtime_probabilities": [0.16495688259601593, 0.022324485704302788, 0.24001085758209229, 0.1455739289522171, 0.07792011648416519, 0.34921371936798096], "decide_seconds": 0.17014330600068206, "probabilities_f64_minus_runtime_max_abs": 1.0146878309802432e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.10230179331729657, "hypatia": 0.09028101573136368, "sagan": 0.03763472655640346, "socrates": 0.4584848289861117, "darwin": 0.27808510579318024, "einstein": 0.03321252961564426}, "confidence": 0.21847787032138655, "choice": "socrates"}}, "usage": {"input_tokens": 295, "output_tokens": 1}, "latency_ms": 173.4807489992818, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 295, "compiled_prompt_sha256": "a5cdaad818d34b9a9d13e3f249d193a4388ffe12a104296b34eeedb316bd586d", "compiled_ids_sha256": "a3ae59e3cdd049d8dd5a4241c2590bc4e237f2e3e70b0bf3f52141729134efee", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 17.375, 16.5, 19.0, 18.5, 16.375], "runtime_probabilities": [0.10230179876089096, 0.09028100967407227, 0.037634726613759995, 0.45848482847213745, 0.2780851125717163, 0.033212531358003616], "decide_seconds": 0.16985723799734842, "probabilities_f64_minus_runtime_max_abs": 6.778536065787932e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.24750782806259655, "darwin": 0.07091218000038109, "sagan": 0.037956554766399746, "ibn_sina": 0.07091218000038109, "einstein": 0.04873718104995752, "socrates": 0.5239740761202839}, "confidence": 0.25716624128395593, "choice": "socrates"}}, "usage": {"input_tokens": 295, "output_tokens": 1}, "latency_ms": 173.5889080009656, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 295, "compiled_prompt_sha256": "e8618ebc3c93376c0d6c137dc53cca0882fcc0f2c62737655841c3f0f9f0e650", "compiled_ids_sha256": "910f450841c9c0d8ba700d9d2116bedbccf5283d2c303b051072c23db9aec7fe", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.625, 17.375, 16.75, 17.375, 17.0, 19.375], "runtime_probabilities": [0.24750782549381256, 0.07091217488050461, 0.03795655444264412, 0.07091217488050461, 0.04873717576265335, 0.5239740610122681], "decide_seconds": 0.16998939699988114, "probabilities_f64_minus_runtime_max_abs": 1.510801583837207e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.2397677076235133, "einstein": 0.02527133045849411, "darwin": 0.1867312784524278, "sagan": 0.03676956885308838, "hypatia": 0.2397677076235133, "socrates": 0.2716924069889632}, "confidence": 0.1256586388848241, "choice": "socrates"}}, "usage": {"input_tokens": 295, "output_tokens": 1}, "latency_ms": 173.4368600009475, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 295, "compiled_prompt_sha256": "8d3c1b75fc680d29a518eb5fa9bbda1574c0e9b4b3b1b2250a450711dc27cc9f", "compiled_ids_sha256": "0a673aa79942148ced5ad64143b19878f6fc0f3e35a82c3eaf5fd01c73e6d1fc", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.375, 16.125, 18.125, 16.5, 18.375, 18.5], "runtime_probabilities": [0.23976768553256989, 0.025271328166127205, 0.1867312788963318, 0.0367695651948452, 0.23976768553256989, 0.2716923952102661], "decide_seconds": 0.169869067001855, "probabilities_f64_minus_runtime_max_abs": 2.209094340788198e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.03985063198402705, "sagan": 0.011417397221333864, "socrates": 0.7063693534251638, "ibn_sina": 0.051169224338560165, "darwin": 0.09559669651545752, "hypatia": 0.09559669651545752}, "confidence": 0.42738388558140095, "choice": "socrates"}}, "usage": {"input_tokens": 295, "output_tokens": 1}, "latency_ms": 174.48689200318768, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 295, "compiled_prompt_sha256": "d2bea664d466f575873c7433bfa2ddd0d814a958f44532e762cb9be6a106ac51", "compiled_ids_sha256": "bc3c3b2e426a7c32f6dbef6abe826301cdfffb463cbcef5e3ce2be6fcadce610", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 15.5, 19.625, 17.0, 17.625, 17.625], "runtime_probabilities": [0.0398506298661232, 0.011417396366596222, 0.7063692808151245, 0.051169224083423615, 0.09559668600559235, 0.09559668600559235], "decide_seconds": 0.17064559099890175, "probabilities_f64_minus_runtime_max_abs": 7.261003931446908e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.5837156465950228, "ibn_sina": 0.025646675619389863, "hypatia": 0.31244047084533805, "einstein": 0.022633111795703414, "sagan": 0.022633111795703414, "darwin": 0.03293098334884241}, "confidence": 0.4108839390755722, "choice": "socrates"}}, "usage": {"input_tokens": 295, "output_tokens": 1}, "latency_ms": 174.26248300034786, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 295, "compiled_prompt_sha256": "612b973b2241207c94ab421257e98cc6d1faba4ecac98ae2e27795cd7dcb2c6d", "compiled_ids_sha256": "27119e9fb6228f170ae6889611be15eac66a0a3b73c27ea4e28fbd8e15a56251", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.5, 16.375, 18.875, 16.25, 16.25, 16.625], "runtime_probabilities": [0.5837156772613525, 0.025646675378084183, 0.3124404847621918, 0.022633111104369164, 0.022633111104369164, 0.03293098509311676], "decide_seconds": 0.1706474509992404, "probabilities_f64_minus_runtime_max_abs": 3.0666329697837114e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:41Z"} {"id": "c-darwin-01-main-02", "uid": "c008", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "prompt_sha256_by_order": ["99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce"], "request_sha256": "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "compiled_prompt_sha256_by_order": ["a94c81da04c23301082578b21ecddd1e9b9cf59dce663b78a8e4f45435535a31", "116fca7c5b486793578615fa23511bb9dfc0faca55af0b49790f42fdf6a70204", "ce5c61932522695a78d607f502da7286e6efaeb83ce3fee25065467dd752e5e5", "4891f1a20d154ecbf966cb82a1e1f02fde085fb655fafbc89560787dec24a9b0", "a93b82d3dcdd041fd213f1424827f4178fade3c3c3a8c7f316740083c0934f77", "ec6e43a445db0ba3c7b900c68a8a7ba2f5cdc80561b30f7839eec9c7e6dd58d6"], "compiled_tokens_by_order": [372, 372, 372, 372, 372, 372], "choice": "darwin", "probs": {"darwin": 0.9978723304944916, "einstein": 0.00029541484471244694, "hypatia": 0.000754368233213001, "ibn_sina": 0.00029541484471244694, "sagan": 0.00048705673815798654, "socrates": 0.00029541484471244694}, "confidence": 0.9896936488006758, "seconds": 0.19365219566680025, "seconds_all_orders": 1.1619131740008015, "seconds_by_order": [0.19411130899970885, 0.19336191500042332, 0.1934051839998574, 0.19350433300132863, 0.19385496100221644, 0.19367547199726687], "argmax_stable": true, "spread": {"darwin": 0.006106953693718498, "einstein": 0.0017936438411139388, "hypatia": 0.003600055446970878, "ibn_sina": 0.0017999378233383525, "sagan": 0.00043288595222796373, "socrates": 0.0013048412073041588}, "spread_2dp": {"darwin": 0.010000000000000009, "einstein": 0.0, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.0, "socrates": 0.0}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.9978723304944916, "einstein": 0.00029541484471244694, "hypatia": 0.000754368233213001, "ibn_sina": 0.00029541484471244694, "sagan": 0.00048705673815798654, "socrates": 0.00029541484471244694}, {"ibn_sina": 0.0019232375859364497, "hypatia": 0.0007075195683542148, "sagan": 0.0002949380362262045, "socrates": 0.0002949380362262045, "darwin": 0.9962617343992635, "einstein": 0.0005176323739935254}, {"hypatia": 0.00045811419393306283, "darwin": 0.9991083245825011, "sagan": 8.474260491114423e-05, "ibn_sina": 0.0001232997625980971, "einstein": 0.0001232997625980971, "socrates": 0.00010221909345855314}, {"ibn_sina": 0.0009090905158255975, "einstein": 0.00027725694252545577, "darwin": 0.9969388036671816, "sagan": 0.00042942395309008836, "hypatia": 0.001242579408991684, "socrates": 0.00020284551238558424}, {"einstein": 0.001916943603712036, "sagan": 0.00033311484938914785, "socrates": 0.00031293244093330905, "ibn_sina": 0.0003774685762789017, "darwin": 0.9930013708887826, "hypatia": 0.0040581696409039405}, {"socrates": 0.001407060300762712, "ibn_sina": 0.0004031295256400588, "hypatia": 0.0006646482237700176, "einstein": 0.0007531451065986093, "sagan": 0.000517628557139108, "darwin": 0.9962543882860896}], "prompt_tokens": 372, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.9978723304944916, "einstein": 0.00029541484471244694, "hypatia": 0.000754368233213001, "ibn_sina": 0.00029541484471244694, "sagan": 0.00048705673815798654, "socrates": 0.00029541484471244694}, "confidence": 0.9896936488006758, "choice": "darwin"}}, "usage": {"input_tokens": 372, "output_tokens": 1}, "latency_ms": 194.08337899949402, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 372, "compiled_prompt_sha256": "a94c81da04c23301082578b21ecddd1e9b9cf59dce663b78a8e4f45435535a31", "compiled_ids_sha256": "6aa68021cc66abd7749c997fa9d95159d904c84b3398aad7280decc31822aa57", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.875, 12.75, 13.6875, 12.75, 13.25, 12.75], "runtime_probabilities": [0.9978723526000977, 0.00029541488038375974, 0.0007543682586401701, 0.00029541488038375974, 0.0004870567936450243, 0.00029541488038375974], "decide_seconds": 0.189640012999007, "probabilities_f64_minus_runtime_max_abs": 2.2105606012345902e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0019232375859364497, "hypatia": 0.0007075195683542148, "sagan": 0.0002949380362262045, "socrates": 0.0002949380362262045, "darwin": 0.9962617343992635, "einstein": 0.0005176323739935254}, "confidence": 0.9834786082695527, "choice": "darwin"}}, "usage": {"input_tokens": 372, "output_tokens": 1}, "latency_ms": 193.3367849997012, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 372, "compiled_prompt_sha256": "116fca7c5b486793578615fa23511bb9dfc0faca55af0b49790f42fdf6a70204", "compiled_ids_sha256": "0638fc3f06490b8f739806553d9dd009120b9c42242235454840e71c731ba0d5", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [13.875, 12.875, 12.0, 12.0, 20.125, 12.5625], "runtime_probabilities": [0.001923237694427371, 0.0007075195899233222, 0.00029493807232938707, 0.00029493807232938707, 0.9962617754936218, 0.000517632404807955], "decide_seconds": 0.18908414800171158, "probabilities_f64_minus_runtime_max_abs": 4.109435836063824e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.00045811419393306283, "darwin": 0.9991083245825011, "sagan": 8.474260491114423e-05, "ibn_sina": 0.0001232997625980971, "einstein": 0.0001232997625980971, "socrates": 0.00010221909345855314}, "confidence": 0.9953303879371015, "choice": "darwin"}}, "usage": {"input_tokens": 372, "output_tokens": 1}, "latency_ms": 193.3804340005736, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 372, "compiled_prompt_sha256": "ce5c61932522695a78d607f502da7286e6efaeb83ce3fee25065467dd752e5e5", "compiled_ids_sha256": "29a415f9cc2e86b9c574bc0d3b6b13218727624569c4575ca1e736cff56928a4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [13.3125, 21.0, 11.625, 12.0, 12.0, 11.8125], "runtime_probabilities": [0.00045811422751285136, 0.9991083741188049, 8.474260539514944e-05, 0.00012329977471381426, 0.00012329977471381426, 0.00010221909906249493], "decide_seconds": 0.1891181780010811, "probabilities_f64_minus_runtime_max_abs": 4.9536303836639206e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0009090905158255975, "einstein": 0.00027725694252545577, "darwin": 0.9969388036671816, "sagan": 0.00042942395309008836, "hypatia": 0.001242579408991684, "socrates": 0.00020284551238558424}, "confidence": 0.9860128890479594, "choice": "darwin"}}, "usage": {"input_tokens": 372, "output_tokens": 1}, "latency_ms": 193.48005299980287, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 372, "compiled_prompt_sha256": "4891f1a20d154ecbf966cb82a1e1f02fde085fb655fafbc89560787dec24a9b0", "compiled_ids_sha256": "263af896e08189ae7ec587ee49884bd10bbb6f703096e4710444629843c716df", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [13.375, 12.1875, 20.375, 12.625, 13.6875, 11.875], "runtime_probabilities": [0.0009090905077755451, 0.00027725694235414267, 0.9969388246536255, 0.00042942396248690784, 0.0012425794266164303, 0.00020284552010707557], "decide_seconds": 0.18917357700047432, "probabilities_f64_minus_runtime_max_abs": 2.0986443916015673e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.001916943603712036, "sagan": 0.00033311484938914785, "socrates": 0.00031293244093330905, "ibn_sina": 0.0003774685762789017, "darwin": 0.9930013708887826, "hypatia": 0.0040581696409039405}, "confidence": 0.9723821375460111, "choice": "darwin"}}, "usage": {"input_tokens": 372, "output_tokens": 1}, "latency_ms": 193.8302210000984, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 372, "compiled_prompt_sha256": "a93b82d3dcdd041fd213f1424827f4178fade3c3c3a8c7f316740083c0934f77", "compiled_ids_sha256": "86716e25889c79faa630702c84b7931c01c704a2c123e2e1593818c54db5555e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.125, 12.375, 12.3125, 12.5, 20.375, 14.875], "runtime_probabilities": [0.0019169434672221541, 0.0003331148182041943, 0.0003129324468318373, 0.0003774685610551387, 0.9930013418197632, 0.004058169666677713], "decide_seconds": 0.18936577500062413, "probabilities_f64_minus_runtime_max_abs": 2.9069019413796582e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.001407060300762712, "ibn_sina": 0.0004031295256400588, "hypatia": 0.0006646482237700176, "einstein": 0.0007531451065986093, "sagan": 0.000517628557139108, "darwin": 0.9962543882860896}, "confidence": 0.9830758738545685, "choice": "darwin"}}, "usage": {"input_tokens": 372, "output_tokens": 1}, "latency_ms": 193.65004299834254, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 372, "compiled_prompt_sha256": "ec6e43a445db0ba3c7b900c68a8a7ba2f5cdc80561b30f7839eec9c7e6dd58d6", "compiled_ids_sha256": "7b6a22dceaeea48800c4fd7032f2cb4080502ef42f714548914c636bf2e5fed2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [13.9375, 12.6875, 13.1875, 13.3125, 12.9375, 20.5], "runtime_probabilities": [0.00140706030651927, 0.0004031294956803322, 0.0006646481924690306, 0.0007531450828537345, 0.0005176285048946738, 0.9962543249130249], "decide_seconds": 0.1893111460012733, "probabilities_f64_minus_runtime_max_abs": 6.337306468484627e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:42Z"} {"id": "c-darwin-01-main-02", "uid": "c009", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "prompt_sha256_by_order": ["e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1"], "request_sha256": "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "compiled_prompt_sha256_by_order": ["67d1a2d7e7aebb688c009d8c350b3caab3a77e09611db29376cf5a84ebb7f2f5", "6b489c00d800c69dea62a687b738e00186272d8e8645d2b0ffc3bf9e329b17f2", "3152d2d4b8bd9212d970810c5ccb53ed516fad4fa875eec7f36807d575252e9e", "a1d8850e960e60d0520aecdda88b8f09a7c59bf25dd5970dbaf74baa3604f978", "43530d513862b058270391af2e61b92fc71bee46b380f65227e9983640227649", "10e933efb739aa29898f0d6ec85216a1e3a887f656ea440e3dfb73055489427a"], "compiled_tokens_by_order": [407, 407, 407, 407, 407, 407], "choice": "darwin", "probs": {"darwin": 0.9915034359239406, "einstein": 0.0007041403466425721, "hypatia": 0.001315507356079909, "ibn_sina": 0.0040520479230729475, "sagan": 0.0014003502912127467, "socrates": 0.0010245181590512265}, "confidence": 0.9660262746491867, "seconds": 0.2123193208323452, "seconds_all_orders": 1.2739159249940712, "seconds_by_order": [0.2123460669972701, 0.21230311799808987, 0.21173801199984155, 0.21293626199985738, 0.21192755100128124, 0.2126649149977311], "argmax_stable": true, "spread": {"darwin": 0.017533962384882584, "einstein": 0.001911031641015996, "hypatia": 0.0009807940468435565, "ibn_sina": 0.01523718576063763, "sagan": 0.0023649856213075037, "socrates": 0.003332396016485347}, "spread_2dp": {"darwin": 0.020000000000000018, "einstein": 0.0, "hypatia": 0.0, "ibn_sina": 0.02, "sagan": 0.0, "socrates": 0.0}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.9915034359239406, "einstein": 0.0007041403466425721, "hypatia": 0.001315507356079909, "ibn_sina": 0.0040520479230729475, "sagan": 0.0014003502912127467, "socrates": 0.0010245181590512265}, {"ibn_sina": 0.0178523577482962, "hypatia": 0.0017676227879891704, "sagan": 0.002132160076438386, "socrates": 0.0016605279371638018, "darwin": 0.9747057067868411, "einstein": 0.0018816246632714102}, {"hypatia": 0.0020390105793771495, "darwin": 0.9922396691717237, "sagan": 0.0011617932075737483, "ibn_sina": 0.0026181414088154577, "einstein": 0.0012367224318706614, "socrates": 0.0007046632006393619}, {"ibn_sina": 0.006241994376231868, "einstein": 0.001578223335101966, "darwin": 0.9861413909900334, "sagan": 0.002948509365729448, "hypatia": 0.0022963014029234656, "socrates": 0.0007935805299798827}, {"einstein": 0.002615171987658568, "sagan": 0.0005835237444219439, "socrates": 0.0009037792553331469, "ibn_sina": 0.002615171987658568, "darwin": 0.9911142992981091, "hypatia": 0.0021680537268186674}, {"socrates": 0.004037059217124709, "ibn_sina": 0.003346839642814955, "hypatia": 0.00179143416823712, "einstein": 0.0020299608565089488, "sagan": 0.0009588856119883298, "darwin": 0.9878358205033259}], "prompt_tokens": 407, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.9915034359239406, "einstein": 0.0007041403466425721, "hypatia": 0.001315507356079909, "ibn_sina": 0.0040520479230729475, "sagan": 0.0014003502912127467, "socrates": 0.0010245181590512265}, "confidence": 0.9660262746491867, "choice": "darwin"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 212.3172270003124, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "67d1a2d7e7aebb688c009d8c350b3caab3a77e09611db29376cf5a84ebb7f2f5", "compiled_ids_sha256": "b6064067571c6aaaef5c497a87a07eae439d78eab1c8465b4f06fd5fd0319cf5", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [21.0, 13.75, 14.375, 15.5, 14.4375, 14.125], "runtime_probabilities": [0.9915033578872681, 0.0007041402859613299, 0.0013155072228983045, 0.0040520476177334785, 0.0014003501273691654, 0.0010245180455967784], "decide_seconds": 0.20741037499828963, "probabilities_f64_minus_runtime_max_abs": 7.80366725683379e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0178523577482962, "hypatia": 0.0017676227879891704, "sagan": 0.002132160076438386, "socrates": 0.0016605279371638018, "darwin": 0.9747057067868411, "einstein": 0.0018816246632714102}, "confidence": 0.9198593643502984, "choice": "darwin"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 212.27717799774837, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "6b489c00d800c69dea62a687b738e00186272d8e8645d2b0ffc3bf9e329b17f2", "compiled_ids_sha256": "6cf73cd9168f1c04c249ee559e887ecf70908f26037ca053437bb6e75d74ce6a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 14.0625, 14.25, 14.0, 20.375, 14.125], "runtime_probabilities": [0.017852356657385826, 0.0017676225397735834, 0.0021321598906069994, 0.0016605278942734003, 0.9747056365013123, 0.0018816245719790459], "decide_seconds": 0.20750537400090252, "probabilities_f64_minus_runtime_max_abs": 7.028552884147388e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0020390105793771495, "darwin": 0.9922396691717237, "sagan": 0.0011617932075737483, "ibn_sina": 0.0026181414088154577, "einstein": 0.0012367224318706614, "socrates": 0.0007046632006393619}, "confidence": 0.9680907587953405, "choice": "darwin"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 211.71234199937317, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "3152d2d4b8bd9212d970810c5ccb53ed516fad4fa875eec7f36807d575252e9e", "compiled_ids_sha256": "8ad081105ad1e05226a47cacb407c9b3202a64ae11e9af9a01fb847c4e7ea348", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0625, 21.25, 14.5, 15.3125, 14.5625, 14.0], "runtime_probabilities": [0.002039010636508465, 0.9922397136688232, 0.0011617932468652725, 0.0026181417051702738, 0.0012367225717753172, 0.0007046632235869765], "decide_seconds": 0.20700230800139252, "probabilities_f64_minus_runtime_max_abs": 4.4497099560558695e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.006241994376231868, "einstein": 0.001578223335101966, "darwin": 0.9861413909900334, "sagan": 0.002948509365729448, "hypatia": 0.0022963014029234656, "socrates": 0.0007935805299798827}, "confidence": 0.9484142564074591, "choice": "darwin"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 212.91021299839485, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "a1d8850e960e60d0520aecdda88b8f09a7c59bf25dd5970dbaf74baa3604f978", "compiled_ids_sha256": "9e5b8395a56beaa0a1896dfc8384c81c6512380e4eb2cdf6f58faa5b4b8e1561", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5625, 14.1875, 20.625, 14.8125, 14.5625, 13.5], "runtime_probabilities": [0.006241994444280863, 0.0015782234258949757, 0.9861414432525635, 0.002948509529232979, 0.0022963013034313917, 0.0007935805479064584], "decide_seconds": 0.20791298100084532, "probabilities_f64_minus_runtime_max_abs": 5.226253008139281e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.002615171987658568, "sagan": 0.0005835237444219439, "socrates": 0.0009037792553331469, "ibn_sina": 0.002615171987658568, "darwin": 0.9911142992981091, "hypatia": 0.0021680537268186674}, "confidence": 0.9643220397975107, "choice": "darwin"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 211.90219100026297, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "43530d513862b058270391af2e61b92fc71bee46b380f65227e9983640227649", "compiled_ids_sha256": "c7bb9afc45b9b132c444c5d82278775f4ca5ed974d8270acb37de6e52bb95fa4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.4375, 12.9375, 13.375, 14.4375, 20.375, 14.25], "runtime_probabilities": [0.002615172415971756, 0.0005835238262079656, 0.0009037793497554958, 0.002615172415971756, 0.9911143779754639, 0.0021680539939552546], "decide_seconds": 0.20716242699927534, "probabilities_f64_minus_runtime_max_abs": 7.867735474231807e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.004037059217124709, "ibn_sina": 0.003346839642814955, "hypatia": 0.00179143416823712, "einstein": 0.0020299608565089488, "sagan": 0.0009588856119883298, "darwin": 0.9878358205033259}, "confidence": 0.953119363314453, "choice": "darwin"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 212.63930500208517, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "10e933efb739aa29898f0d6ec85216a1e3a887f656ea440e3dfb73055489427a", "compiled_ids_sha256": "b44559e03bdf5fd54bda7ce3b93240c4bfc7ef936bc36685d683b0f19bbde4d8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.25, 15.0625, 14.4375, 14.5625, 13.8125, 20.75], "runtime_probabilities": [0.004037059377878904, 0.003346839686855674, 0.0017914341297000647, 0.0020299607422202826, 0.0009588855900801718, 0.9878358244895935], "decide_seconds": 0.20791182100219885, "probabilities_f64_minus_runtime_max_abs": 3.986267649302988e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:43Z"} {"id": "c-darwin-01-main-02", "uid": "c010", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "prompt_sha256_by_order": ["b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285"], "request_sha256": "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "compiled_prompt_sha256_by_order": ["845aaf3e77ec7fd9ccefeb0369bad6914f74a0a4e47c5038bcac81310111f0a8", "c0f56b4a3afc646003c0ad4d6a0bf9b4c61e37dcea864a1b30612c7c4b598c46", "dad99b7d8ff8c4ed85f54efc8b59d31ce90010731bc3f3d998d8822eb7f3fff5", "711cb74e2e74eee811b11c2a3657aaad3418f80fda1edf8aafc2f409485d5561", "c6a195b28e61cccd7241e6669c72ef9c4393c217ebfc859cb0ebb54828c32650", "60b895f7526886c0f0e93a5b2c1e6b89b928d31850472819195ecd2731436637"], "compiled_tokens_by_order": [401, 401, 401, 401, 401, 401], "choice": "darwin", "probs": {"darwin": 0.4165451456843789, "einstein": 0.03419210771029396, "hypatia": 0.09294378506560648, "ibn_sina": 0.4165451456843789, "sagan": 0.03419210771029396, "socrates": 0.005581708145047879}, "confidence": 0.3245701872604473, "seconds": 0.21175268483299683, "seconds_all_orders": 1.270516108997981, "seconds_by_order": [0.2114434240029368, 0.2110502469986386, 0.21220359799917787, 0.21139668400064693, 0.21217171799798962, 0.21225043799859122], "argmax_stable": false, "spread": {"darwin": 0.5781826646746877, "einstein": 0.06673842188347463, "hypatia": 0.08884279892322533, "ibn_sina": 0.5058606761210361, "sagan": 0.031199506398249494, "socrates": 0.016736080461480365}, "spread_2dp": {"darwin": 0.5800000000000001, "einstein": 0.07, "hypatia": 0.09000000000000001, "ibn_sina": 0.51, "sagan": 0.03, "socrates": 0.02}, "choices_by_order": ["darwin", "darwin", "ibn_sina", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.4165451456843789, "einstein": 0.03419210771029396, "hypatia": 0.09294378506560648, "ibn_sina": 0.4165451456843789, "sagan": 0.03419210771029396, "socrates": 0.005581708145047879}, {"ibn_sina": 0.0728716976980275, "hypatia": 0.011895977634104438, "sagan": 0.006778123739295909, "socrates": 0.007215275162340096, "darwin": 0.8877590171124291, "einstein": 0.013479908653803009}, {"hypatia": 0.041896603340519166, "darwin": 0.30957635243774145, "sagan": 0.006839432482403277, "ibn_sina": 0.5783647689584234, "einstein": 0.06095919816211964, "socrates": 0.0023636446187931533}, {"ibn_sina": 0.12935114954563864, "einstein": 0.008802456195697517, "darwin": 0.7443644713197495, "sagan": 0.012031531103799722, "hypatia": 0.10073877655732977, "socrates": 0.00471161527778489}, {"einstein": 0.020772770389908, "sagan": 0.002992601312044469, "socrates": 0.005252178869129145, "ibn_sina": 0.07250409283738726, "darwin": 0.8832806731177544, "hypatia": 0.015197683473776721}, {"socrates": 0.01909972508027352, "ibn_sina": 0.2636635718041221, "hypatia": 0.06666459092338999, "einstein": 0.07554087807917215, "sagan": 0.016855448223558694, "darwin": 0.5581757858894836}], "prompt_tokens": 401, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.4165451456843789, "einstein": 0.03419210771029396, "hypatia": 0.09294378506560648, "ibn_sina": 0.4165451456843789, "sagan": 0.03419210771029396, "socrates": 0.005581708145047879}, "confidence": 0.3245701872604473, "choice": "darwin"}}, "usage": {"input_tokens": 401, "output_tokens": 1}, "latency_ms": 211.41161400009878, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 401, "compiled_prompt_sha256": "845aaf3e77ec7fd9ccefeb0369bad6914f74a0a4e47c5038bcac81310111f0a8", "compiled_ids_sha256": "2c82207f9a93c3dc26bdba65b250a344432061dc02ca1018c708ad8067c3121f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.25, 16.75, 17.75, 19.25, 16.75, 14.9375], "runtime_probabilities": [0.41654515266418457, 0.03419210761785507, 0.09294378757476807, 0.41654515266418457, 0.03419210761785507, 0.005581708624958992], "decide_seconds": 0.20663984100247035, "probabilities_f64_minus_runtime_max_abs": 6.979805677875817e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0728716976980275, "hypatia": 0.011895977634104438, "sagan": 0.006778123739295909, "socrates": 0.007215275162340096, "darwin": 0.8877590171124291, "einstein": 0.013479908653803009}, "confidence": 0.733920993101704, "choice": "darwin"}}, "usage": {"input_tokens": 401, "output_tokens": 1}, "latency_ms": 211.02213699850836, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 401, "compiled_prompt_sha256": "c0f56b4a3afc646003c0ad4d6a0bf9b4c61e37dcea864a1b30612c7c4b598c46", "compiled_ids_sha256": "885b49844a750f95496e2fcd51c3a0a038a50cfd500fcde03ed2de46a8309fec", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 15.3125, 14.75, 14.8125, 19.625, 15.4375], "runtime_probabilities": [0.07287169992923737, 0.011895977891981602, 0.006778123322874308, 0.0072152758948504925, 0.8877590298652649, 0.013479907996952534], "decide_seconds": 0.2063029339988134, "probabilities_f64_minus_runtime_max_abs": 1.2752835787388506e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.041896603340519166, "darwin": 0.30957635243774145, "sagan": 0.006839432482403277, "ibn_sina": 0.5783647689584234, "einstein": 0.06095919816211964, "socrates": 0.0023636446187931533}, "confidence": 0.42429630435024335, "choice": "ibn_sina"}}, "usage": {"input_tokens": 401, "output_tokens": 1}, "latency_ms": 212.17492799769389, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 401, "compiled_prompt_sha256": "dad99b7d8ff8c4ed85f54efc8b59d31ce90010731bc3f3d998d8822eb7f3fff5", "compiled_ids_sha256": "656aeb0145a2ecafd9b7f8fad09ab24bea6eef2b61e89c8eb9f4ca7b83b8f6c2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 18.875, 15.0625, 19.5, 17.25, 14.0], "runtime_probabilities": [0.041896604001522064, 0.3095763623714447, 0.006839431822299957, 0.5783647298812866, 0.06095919385552406, 0.002363644540309906], "decide_seconds": 0.20727854700089665, "probabilities_f64_minus_runtime_max_abs": 3.907713674866642e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.12935114954563864, "einstein": 0.008802456195697517, "darwin": 0.7443644713197495, "sagan": 0.012031531103799722, "hypatia": 0.10073877655732977, "socrates": 0.00471161527778489}, "confidence": 0.5336371427185702, "choice": "darwin"}}, "usage": {"input_tokens": 401, "output_tokens": 1}, "latency_ms": 211.3693940009398, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 401, "compiled_prompt_sha256": "711cb74e2e74eee811b11c2a3657aaad3418f80fda1edf8aafc2f409485d5561", "compiled_ids_sha256": "48f156de0c08b8519effc342a4da6257c28bcace39978846d0ed7b62aa58fdd6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 15.0625, 19.5, 15.375, 17.5, 14.4375], "runtime_probabilities": [0.12935113906860352, 0.008802455849945545, 0.7443645000457764, 0.01203153096139431, 0.1007387787103653, 0.004711615387350321], "decide_seconds": 0.20665169100175262, "probabilities_f64_minus_runtime_max_abs": 2.872602689052428e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.020772770389908, "sagan": 0.002992601312044469, "socrates": 0.005252178869129145, "ibn_sina": 0.07250409283738726, "darwin": 0.8832806731177544, "hypatia": 0.015197683473776721}, "confidence": 0.7271124718372661, "choice": "darwin"}}, "usage": {"input_tokens": 401, "output_tokens": 1}, "latency_ms": 212.14343899919186, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 401, "compiled_prompt_sha256": "c6a195b28e61cccd7241e6669c72ef9c4393c217ebfc859cb0ebb54828c32650", "compiled_ids_sha256": "b062188d673891c8e1dfc1c7ebc45971b69840fbb55bfe5feacc5b0bcedeb8f9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 14.1875, 14.75, 17.375, 19.875, 15.8125], "runtime_probabilities": [0.020772771909832954, 0.0029926015995442867, 0.005252178758382797, 0.07250409573316574, 0.8832806944847107, 0.015197684057056904], "decide_seconds": 0.2074440249998588, "probabilities_f64_minus_runtime_max_abs": 2.136695631538288e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.01909972508027352, "ibn_sina": 0.2636635718041221, "hypatia": 0.06666459092338999, "einstein": 0.07554087807917215, "sagan": 0.016855448223558694, "darwin": 0.5581757858894836}, "confidence": 0.3319255793279914, "choice": "darwin"}}, "usage": {"input_tokens": 401, "output_tokens": 1}, "latency_ms": 212.22173800197197, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 401, "compiled_prompt_sha256": "60b895f7526886c0f0e93a5b2c1e6b89b928d31850472819195ecd2731436637", "compiled_ids_sha256": "5b63a165937f644ee727fd04f307c0fc69c779463d881b0bf2751da3d83761c6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 18.75, 17.375, 17.5, 16.0, 19.5], "runtime_probabilities": [0.019099723547697067, 0.26366356015205383, 0.06666459143161774, 0.07554087042808533, 0.016855448484420776, 0.5581757426261902], "decide_seconds": 0.20727966699996614, "probabilities_f64_minus_runtime_max_abs": 4.326329339221502e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:44Z"} {"id": "c-darwin-01-main-03", "uid": "c011", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "prompt_sha256_by_order": ["c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5"], "request_sha256": "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "compiled_prompt_sha256_by_order": ["67ed85f77269ae70491a1864dafa1c40fd5c3c6a2590a97dc9bb1b6af2efd7d9", "d52ddc7906085a25361cdef95ad18623fd050948d5f3ee668e7c6773611100d9", "4dd4b32c9d73f073a47034e6e6e747f7eb03445fd051cbc33402493a06c5475a", "28b60c3fc9331143d1ccf39bc628a8f9092336156e1fb92ca3dadee0363bf83a", "2db636df227dd59cce1576cba21f762a6d76822ed2d9d94692fe9f89b0d7e44a", "d59853cbf40844109510a84c212219d3b41cec000aa6123862bd5ad5559439a5"], "compiled_tokens_by_order": [402, 402, 402, 402, 402, 402], "choice": "darwin", "probs": {"darwin": 0.9937645278353487, "einstein": 0.0003548714332601708, "hypatia": 0.0021738510782050337, "ibn_sina": 0.0009646405684703391, "sagan": 0.002314052427261873, "socrates": 0.00042805665745379665}, "confidence": 0.9740896027356354, "seconds": 0.21252780383292702, "seconds_all_orders": 1.2751668229975621, "seconds_by_order": [0.21265884500098764, 0.2126522250000562, 0.2130756709993875, 0.21178190099817584, 0.21273672399911447, 0.21226145699984045], "argmax_stable": true, "spread": {"darwin": 0.004948358641674422, "einstein": 0.00143838028266286, "hypatia": 0.005076142413896679, "ibn_sina": 0.0031498311810852844, "sagan": 0.001767141553943809, "socrates": 0.0013921631133569817}, "spread_2dp": {"darwin": 0.0, "einstein": 0.0, "hypatia": 0.01, "ibn_sina": 0.0, "sagan": 0.0, "socrates": 0.0}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.9937645278353487, "einstein": 0.0003548714332601708, "hypatia": 0.0021738510782050337, "ibn_sina": 0.0009646405684703391, "sagan": 0.002314052427261873, "socrates": 0.00042805665745379665}, {"ibn_sina": 0.004051514813021699, "hypatia": 0.001586595997343697, "sagan": 0.0009040151490732677, "socrates": 0.0008492436400206993, "darwin": 0.9913729881953913, "einstein": 0.0012356422051492276}, {"hypatia": 0.002313007352961175, "darwin": 0.993315723064657, "sagan": 0.001163053137147857, "ibn_sina": 0.0015897051554549016, "einstein": 0.001163053137147857, "socrates": 0.0004554581526312035}, {"ibn_sina": 0.0014900774730091228, "einstein": 0.00037675058599338464, "darwin": 0.991112563759904, "sagan": 0.0023078771375087143, "hypatia": 0.004311682132400904, "socrates": 0.00040104891118401446}, {"einstein": 0.0017932517159230308, "sagan": 0.000546910873318064, "socrates": 0.0008470721312474715, "ibn_sina": 0.0013119709820046729, "darwin": 0.9888380558862664, "hypatia": 0.006662738411240376}, {"socrates": 0.0017932120245409962, "ibn_sina": 0.0009016836319364148, "hypatia": 0.004874455660884045, "einstein": 0.0015825040573348805, "sagan": 0.0020319754316294614, "darwin": 0.9888161691936743}], "prompt_tokens": 402, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.9937645278353487, "einstein": 0.0003548714332601708, "hypatia": 0.0021738510782050337, "ibn_sina": 0.0009646405684703391, "sagan": 0.002314052427261873, "socrates": 0.00042805665745379665}, "confidence": 0.9740896027356354, "choice": "darwin"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 212.62903499882668, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 402, "compiled_prompt_sha256": "67ed85f77269ae70491a1864dafa1c40fd5c3c6a2590a97dc9bb1b6af2efd7d9", "compiled_ids_sha256": "f12e7fc6a162cdd7569eba4da42ac19c958d6ecfbd2351179f1c50f8bd12014a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [21.0, 13.0625, 14.875, 14.0625, 14.9375, 13.25], "runtime_probabilities": [0.9937645196914673, 0.00035487141576595604, 0.0021738512441515923, 0.0009646405233070254, 0.0023140523117035627, 0.0004280566645320505], "decide_seconds": 0.20762608300356078, "probabilities_f64_minus_runtime_max_abs": 8.143881458089197e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.004051514813021699, "hypatia": 0.001586595997343697, "sagan": 0.0009040151490732677, "socrates": 0.0008492436400206993, "darwin": 0.9913729881953913, "einstein": 0.0012356422051492276}, "confidence": 0.9655362616181316, "choice": "darwin"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 212.6270549997571, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 402, "compiled_prompt_sha256": "d52ddc7906085a25361cdef95ad18623fd050948d5f3ee668e7c6773611100d9", "compiled_ids_sha256": "9278fdf813c8a5b0b95054149ed642ad320e34842969e3be93fb44b675f2832e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.875, 13.9375, 13.375, 13.3125, 20.375, 13.6875], "runtime_probabilities": [0.0040515149012207985, 0.0015865961322560906, 0.0009040152071975172, 0.0008492436609230936, 0.9913730621337891, 0.001235642354004085], "decide_seconds": 0.20777425200139987, "probabilities_f64_minus_runtime_max_abs": 7.393839773772726e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.002313007352961175, "darwin": 0.993315723064657, "sagan": 0.001163053137147857, "ibn_sina": 0.0015897051554549016, "einstein": 0.001163053137147857, "socrates": 0.0004554581526312035}, "confidence": 0.9720020440559602, "choice": "darwin"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 213.05056199707906, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 402, "compiled_prompt_sha256": "4dd4b32c9d73f073a47034e6e6e747f7eb03445fd051cbc33402493a06c5475a", "compiled_ids_sha256": "2ba8d8484869807159429752d76b9332b0d3916ce4545ab7a943d6a6eb7d05d1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0625, 21.125, 14.375, 14.6875, 14.375, 13.4375], "runtime_probabilities": [0.002313007367774844, 0.9933157563209534, 0.0011630530934780836, 0.0015897052362561226, 0.0011630530934780836, 0.0004554581537377089], "decide_seconds": 0.2080903300011414, "probabilities_f64_minus_runtime_max_abs": 3.325629638339933e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0014900774730091228, "einstein": 0.00037675058599338464, "darwin": 0.991112563759904, "sagan": 0.0023078771375087143, "hypatia": 0.004311682132400904, "socrates": 0.00040104891118401446}, "confidence": 0.9653139542808852, "choice": "darwin"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 211.75721200052067, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 402, "compiled_prompt_sha256": "28b60c3fc9331143d1ccf39bc628a8f9092336156e1fb92ca3dadee0363bf83a", "compiled_ids_sha256": "f47ed8b1971e6c2f470e0da907f26d2fe1530e352f7329a2b51acf4440bd1ba4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.125, 12.75, 20.625, 14.5625, 15.1875, 12.8125], "runtime_probabilities": [0.0014900773530825973, 0.00037675059866160154, 0.9911125302314758, 0.002307876944541931, 0.004311682190746069, 0.00040104889194481075], "decide_seconds": 0.2069354490013211, "probabilities_f64_minus_runtime_max_abs": 3.3528428144968814e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.0017932517159230308, "sagan": 0.000546910873318064, "socrates": 0.0008470721312474715, "ibn_sina": 0.0013119709820046729, "darwin": 0.9888380558862664, "hypatia": 0.006662738411240376}, "confidence": 0.9583457561650046, "choice": "darwin"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 212.71177399830776, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 402, "compiled_prompt_sha256": "2db636df227dd59cce1576cba21f762a6d76822ed2d9d94692fe9f89b0d7e44a", "compiled_ids_sha256": "431e0ab2d134a50837ac255a8b84534c15d958ed6646d807dc372a5d2b6a1917", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.3125, 13.125, 13.5625, 14.0, 20.625, 15.625], "runtime_probabilities": [0.0017932517221197486, 0.0005469108582474291, 0.0008470721077173948, 0.0013119709910824895, 0.9888380765914917, 0.006662738509476185], "decide_seconds": 0.20792407199769514, "probabilities_f64_minus_runtime_max_abs": 2.070522531205654e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0017932120245409962, "ibn_sina": 0.0009016836319364148, "hypatia": 0.004874455660884045, "einstein": 0.0015825040573348805, "sagan": 0.0020319754316294614, "darwin": 0.9888161691936743}, "confidence": 0.9567274283834822, "choice": "darwin"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 212.23681699848385, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 402, "compiled_prompt_sha256": "d59853cbf40844109510a84c212219d3b41cec000aa6123862bd5ad5559439a5", "compiled_ids_sha256": "b810e433af9e97e7b77d389f176bd2d0fe6591ea3d97424584348614ea4ef64c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.3125, 13.625, 15.3125, 14.1875, 14.4375, 20.625], "runtime_probabilities": [0.0017932121409103274, 0.0009016837575472891, 0.004874456208199263, 0.0015825042501091957, 0.002031975658610463, 0.9888162612915039], "decide_seconds": 0.20730445600202074, "probabilities_f64_minus_runtime_max_abs": 9.209782958485846e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:46Z"} {"id": "c-darwin-01-starter-01", "uid": "c012", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "prompt_sha256_by_order": ["97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1"], "request_sha256": "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "compiled_prompt_sha256_by_order": ["9e3f59abe3f3521d1414d75abb72215d06b06fabcaa7452231957e31bc3997a0", "8d8d52d24ae0d69da1b42ca2bf2d9f23899f777722a968bf55dc653738e402ad", "818369bbaf473d78c1e70dc8e86a92ae4c14cf58b91d017b51948f1c911e0516", "0c3bf7de2e1d0269fd25c830086d150e4bdfa13a669c8e7d6419b0f53713ac09", "244a5833bbffa916ed9d6646487ae937bbd4fe94a9654b6e61b8c6ae32278738", "ef53f77a12c03833d222d2c68857c7aa1bf109028bd05604056b23ae46da6e67"], "compiled_tokens_by_order": [240, 240, 240, 240, 240, 240], "choice": "sagan", "probs": {"darwin": 0.047215667732997714, "einstein": 0.032450822223846175, "hypatia": 0.04166768052783391, "ibn_sina": 0.036771599003697884, "sagan": 0.8369177357453957, "socrates": 0.004976494766228598}, "confidence": 0.617884793778829, "seconds": 0.16077893299916468, "seconds_all_orders": 0.9646735979949881, "seconds_by_order": [0.16175343999930192, 0.16108249600074487, 0.16055295999831287, 0.16021229199759546, 0.16042202099924907, 0.1606503889997839], "argmax_stable": false, "spread": {"darwin": 0.391126562334355, "einstein": 0.22046854026497933, "hypatia": 0.05727212650823224, "ibn_sina": 0.05202150635364908, "sagan": 0.5571689393053841, "socrates": 0.044870841680602326}, "spread_2dp": {"darwin": 0.38999999999999996, "einstein": 0.22, "hypatia": 0.049999999999999996, "ibn_sina": 0.049999999999999996, "sagan": 0.55, "socrates": 0.05}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "darwin", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.047215667732997714, "einstein": 0.032450822223846175, "hypatia": 0.04166768052783391, "ibn_sina": 0.036771599003697884, "sagan": 0.8369177357453957, "socrates": 0.004976494766228598}, {"ibn_sina": 0.018146985472682085, "hypatia": 0.006675902874636568, "sagan": 0.8743707259231942, "socrates": 0.01100670307054144, "darwin": 0.0558966504100261, "einstein": 0.03390303224891957}, {"hypatia": 0.02883553744312323, "darwin": 0.02883553744312323, "sagan": 0.8426979292222592, "ibn_sina": 0.02544727247791838, "einstein": 0.06917285836057151, "socrates": 0.0050108650530044894}, {"ibn_sina": 0.04981076563415169, "einstein": 0.03879266328126356, "darwin": 0.09305876153260777, "sagan": 0.7791714706846057, "hypatia": 0.03423440518872226, "socrates": 0.004931933678649107}, {"einstein": 0.13222937741513155, "sagan": 0.31720178661781007, "socrates": 0.037884350915271675, "ibn_sina": 0.06246073518067628, "darwin": 0.4072951562360296, "hypatia": 0.0429285936350809}, {"socrates": 0.049802775359251435, "ibn_sina": 0.010439228827027207, "hypatia": 0.06394802938286881, "einstein": 0.2529193624888255, "sagan": 0.6067220100403524, "darwin": 0.0161685939016746}], "prompt_tokens": 240, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.047215667732997714, "einstein": 0.032450822223846175, "hypatia": 0.04166768052783391, "ibn_sina": 0.036771599003697884, "sagan": 0.8369177357453957, "socrates": 0.004976494766228598}, "confidence": 0.617884793778829, "choice": "sagan"}}, "usage": {"input_tokens": 240, "output_tokens": 1}, "latency_ms": 161.7294210009277, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 240, "compiled_prompt_sha256": "9e3f59abe3f3521d1414d75abb72215d06b06fabcaa7452231957e31bc3997a0", "compiled_ids_sha256": "79ebf5af4c3e548555cf057c1040dde2bbc5c096909a258efb54835918c41f5f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 16.75, 17.0, 16.875, 20.0, 14.875], "runtime_probabilities": [0.047215670347213745, 0.03245082125067711, 0.04166767746210098, 0.03677159920334816, 0.8369176983833313, 0.004976494703441858], "decide_seconds": 0.15850562600098783, "probabilities_f64_minus_runtime_max_abs": 3.736206444227008e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.018146985472682085, "hypatia": 0.006675902874636568, "sagan": 0.8743707259231942, "socrates": 0.01100670307054144, "darwin": 0.0558966504100261, "einstein": 0.03390303224891957}, "confidence": 0.6935021716099224, "choice": "sagan"}}, "usage": {"input_tokens": 240, "output_tokens": 1}, "latency_ms": 161.06194599706214, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 240, "compiled_prompt_sha256": "8d8d52d24ae0d69da1b42ca2bf2d9f23899f777722a968bf55dc653738e402ad", "compiled_ids_sha256": "0a1d8f77807eb4997b7dfd9a04b1578f84abfe58e2bdcfdff3d88729806a2f3b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 15.25, 20.125, 15.75, 17.375, 16.875], "runtime_probabilities": [0.01814698614180088, 0.006675902754068375, 0.8743707537651062, 0.011006703600287437, 0.05589665099978447, 0.033903032541275024], "decide_seconds": 0.15796042999863857, "probabilities_f64_minus_runtime_max_abs": 2.7841911998116586e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.02883553744312323, "darwin": 0.02883553744312323, "sagan": 0.8426979292222592, "ibn_sina": 0.02544727247791838, "einstein": 0.06917285836057151, "socrates": 0.0050108650530044894}, "confidence": 0.6352943096132517, "choice": "sagan"}}, "usage": {"input_tokens": 240, "output_tokens": 1}, "latency_ms": 160.53287999966415, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 240, "compiled_prompt_sha256": "818369bbaf473d78c1e70dc8e86a92ae4c14cf58b91d017b51948f1c911e0516", "compiled_ids_sha256": "303d303aa6647d263a2d077c744c809437adca7e2487101ea586dcf55344d142", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 16.625, 20.0, 16.5, 17.5, 14.875], "runtime_probabilities": [0.028835536912083626, 0.028835536912083626, 0.8426979184150696, 0.02544727362692356, 0.06917285919189453, 0.005010864697396755], "decide_seconds": 0.15745205400162376, "probabilities_f64_minus_runtime_max_abs": 1.0807189587680455e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.04981076563415169, "einstein": 0.03879266328126356, "darwin": 0.09305876153260777, "sagan": 0.7791714706846057, "hypatia": 0.03423440518872226, "socrates": 0.004931933678649107}, "confidence": 0.5353271188023856, "choice": "sagan"}}, "usage": {"input_tokens": 240, "output_tokens": 1}, "latency_ms": 160.1919929998985, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 240, "compiled_prompt_sha256": "0c3bf7de2e1d0269fd25c830086d150e4bdfa13a669c8e7d6419b0f53713ac09", "compiled_ids_sha256": "45f0ff29b7458d0cd0e0cffa3265d3dfe1509bf4f5af99b940c9a55b9010b315", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 16.625, 17.5, 19.625, 16.5, 14.5625], "runtime_probabilities": [0.04981076344847679, 0.03879265859723091, 0.0930587574839592, 0.7791714668273926, 0.03423440456390381, 0.004931933246552944], "decide_seconds": 0.15727088600033312, "probabilities_f64_minus_runtime_max_abs": 4.684032647861436e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.13222937741513155, "sagan": 0.31720178661781007, "socrates": 0.037884350915271675, "ibn_sina": 0.06246073518067628, "darwin": 0.4072951562360296, "hypatia": 0.0429285936350809}, "confidence": 0.20192736517162224, "choice": "darwin"}}, "usage": {"input_tokens": 240, "output_tokens": 1}, "latency_ms": 160.4017509998812, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 240, "compiled_prompt_sha256": "244a5833bbffa916ed9d6646487ae937bbd4fe94a9654b6e61b8c6ae32278738", "compiled_ids_sha256": "9cac17e1afd0057ff1a4a8f232264d2dd594a839a4280ed64b036ad597fea0fd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.0, 18.875, 16.75, 17.25, 19.125, 16.875], "runtime_probabilities": [0.13222937285900116, 0.31720179319381714, 0.03788435086607933, 0.06246073916554451, 0.4072951674461365, 0.04292859137058258], "decide_seconds": 0.15741930500007584, "probabilities_f64_minus_runtime_max_abs": 1.121010689830726e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.049802775359251435, "ibn_sina": 0.010439228827027207, "hypatia": 0.06394802938286881, "einstein": 0.2529193624888255, "sagan": 0.6067220100403524, "darwin": 0.0161685939016746}, "confidence": 0.39143643339616574, "choice": "sagan"}}, "usage": {"input_tokens": 240, "output_tokens": 1}, "latency_ms": 160.63041899906239, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 240, "compiled_prompt_sha256": "ef53f77a12c03833d222d2c68857c7aa1bf109028bd05604056b23ae46da6e67", "compiled_ids_sha256": "7bbba6d95257e9825d1aeefda8a77a4e758aec3f53b58511dc0facd4f924e1e9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 15.4375, 17.25, 18.625, 19.5, 15.875], "runtime_probabilities": [0.04980277270078659, 0.010439228266477585, 0.06394802778959274, 0.25291934609413147, 0.6067219972610474, 0.016168592497706413], "decide_seconds": 0.1576900830004888, "probabilities_f64_minus_runtime_max_abs": 1.6394694046795166e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:47Z"} {"id": "c-darwin-01-starter-01", "uid": "c013", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "prompt_sha256_by_order": ["b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511"], "request_sha256": "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "compiled_prompt_sha256_by_order": ["4503591e6f0e5386c6e8577056733fa060b0d80509ece075b927a593b418dd73", "c99d99efe7d0d9d15d27bb1cf2e7d32fec3f3339cbfc311efe0bf04fb1140bf1", "486326b0f3d7561f7480e438d8bc8483980de1a4517534b089339a49924bfc19", "be9a9ea0c51aa522dde769357e4eeab2be1586c5acd8003b882c6f54d5df37a0", "23be69ba5cb1f451624fd3b7c40ae82e0b4db49b8ca665f6aeb2c3876ffe12a3", "07e98ff80964b0a6def8b7573c24437abd99e5a530b4fe072fb31b28b2c86dd1"], "compiled_tokens_by_order": [269, 269, 269, 269, 269, 269], "choice": "darwin", "probs": {"darwin": 0.9985719314235767, "einstein": 0.00027771112888757345, "hypatia": 0.00027771112888757345, "ibn_sina": 0.00027771112888757345, "sagan": 0.00040406730827534815, "socrates": 0.0001908678814853671}, "confidence": 0.9927214276544472, "seconds": 0.17192899633240208, "seconds_all_orders": 1.0315739779944124, "seconds_by_order": [0.1722255389977363, 0.17179545299950405, 0.17155943399848184, 0.17256883599839057, 0.17195629100024235, 0.17146842500005732], "argmax_stable": true, "spread": {"darwin": 0.005693902336571521, "einstein": 0.0020376226342441457, "hypatia": 0.0008159750992478376, "ibn_sina": 0.003534112913180431, "sagan": 0.0018857460391392886, "socrates": 0.0015451347429405792}, "spread_2dp": {"darwin": 0.010000000000000009, "einstein": 0.0, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.0, "socrates": 0.0}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.9985719314235767, "einstein": 0.00027771112888757345, "hypatia": 0.00027771112888757345, "ibn_sina": 0.00027771112888757345, "sagan": 0.00040406730827534815, "socrates": 0.0001908678814853671}, {"ibn_sina": 0.003811824042068004, "hypatia": 0.0007051165475020585, "sagan": 0.0007051165475020585, "socrates": 0.0006623956955293726, "darwin": 0.9928780290870052, "einstein": 0.0012375180803932554}, {"hypatia": 0.0004301271897612029, "darwin": 0.998571418523784, "sagan": 0.00015823495020199425, "ibn_sina": 0.0003349833922063276, "einstein": 0.0003565879648331438, "socrates": 0.00014864797921339325}, {"ibn_sina": 0.001591853995082056, "einstein": 0.0005501300201530713, "darwin": 0.9946584099023216, "sagan": 0.002043980989341283, "hypatia": 0.0008004344562464567, "socrates": 0.000355190636855577}, {"einstein": 0.002315333763131719, "sagan": 0.0006633542294695925, "socrates": 0.0005854080528233063, "ibn_sina": 0.0010274231293296043, "darwin": 0.9943147945971103, "hypatia": 0.0010936862281354111}, {"socrates": 0.0016937827221539725, "ibn_sina": 0.0008516874389205978, "hypatia": 0.0005853556456510678, "einstein": 0.0014041951965365058, "sagan": 0.0012391979115676335, "darwin": 0.9942257810851703}], "prompt_tokens": 269, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.9985719314235767, "einstein": 0.00027771112888757345, "hypatia": 0.00027771112888757345, "ibn_sina": 0.00027771112888757345, "sagan": 0.00040406730827534815, "socrates": 0.0001908678814853671}, "confidence": 0.9927214276544472, "choice": "darwin"}}, "usage": {"input_tokens": 269, "output_tokens": 1}, "latency_ms": 172.1986199991079, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 269, "compiled_prompt_sha256": "4503591e6f0e5386c6e8577056733fa060b0d80509ece075b927a593b418dd73", "compiled_ids_sha256": "fc3048d46a00e63f863e4d4d211b3e8a6c309b8b184a6adb0f036ee0575056f9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.75, 12.5625, 12.5625, 12.5625, 12.9375, 12.1875], "runtime_probabilities": [0.9985719919204712, 0.00027771113673225045, 0.00027771113673225045, 0.00027771113673225045, 0.00040406733751296997, 0.0001908678823383525], "decide_seconds": 0.1686829360005504, "probabilities_f64_minus_runtime_max_abs": 6.049689449394435e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.003811824042068004, "hypatia": 0.0007051165475020585, "sagan": 0.0007051165475020585, "socrates": 0.0006623956955293726, "darwin": 0.9928780290870052, "einstein": 0.0012375180803932554}, "confidence": 0.9711487027382025, "choice": "darwin"}}, "usage": {"input_tokens": 269, "output_tokens": 1}, "latency_ms": 171.7721929999243, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 269, "compiled_prompt_sha256": "c99d99efe7d0d9d15d27bb1cf2e7d32fec3f3339cbfc311efe0bf04fb1140bf1", "compiled_ids_sha256": "1922fb71f79b3713cc991563e5534f816800e5294bf27df14f7752e21b49372d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.6875, 13.0, 13.0, 12.9375, 20.25, 13.5625], "runtime_probabilities": [0.0038118246011435986, 0.0007051166030578315, 0.0007051166030578315, 0.0006623957888223231, 0.9928781390190125, 0.0012375182705000043], "decide_seconds": 0.16838129899770138, "probabilities_f64_minus_runtime_max_abs": 1.099320072750487e-07}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0004301271897612029, "darwin": 0.998571418523784, "sagan": 0.00015823495020199425, "ibn_sina": 0.0003349833922063276, "einstein": 0.0003565879648331438, "socrates": 0.00014864797921339325}, "confidence": 0.9927624834390026, "choice": "darwin"}}, "usage": {"input_tokens": 269, "output_tokens": 1}, "latency_ms": 171.5358039982675, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 269, "compiled_prompt_sha256": "486326b0f3d7561f7480e438d8bc8483980de1a4517534b089339a49924bfc19", "compiled_ids_sha256": "b68d59f7901b3a45c5da3d4a2126460edcf292f6c42d7730720029352f40c6c2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [13.375, 21.125, 12.375, 13.125, 13.1875, 12.3125], "runtime_probabilities": [0.0004301271983422339, 0.9985713958740234, 0.00015823495050426573, 0.00033498340053483844, 0.0003565879596862942, 0.0001486479741288349], "decide_seconds": 0.16814478099695407, "probabilities_f64_minus_runtime_max_abs": 2.264976051158385e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.001591853995082056, "einstein": 0.0005501300201530713, "darwin": 0.9946584099023216, "sagan": 0.002043980989341283, "hypatia": 0.0008004344562464567, "socrates": 0.000355190636855577}, "confidence": 0.9771738356086128, "choice": "darwin"}}, "usage": {"input_tokens": 269, "output_tokens": 1}, "latency_ms": 172.54583600151818, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 269, "compiled_prompt_sha256": "be9a9ea0c51aa522dde769357e4eeab2be1586c5acd8003b882c6f54d5df37a0", "compiled_ids_sha256": "ca43b1caf0e65a626ec0916d80015aa8d074aef928db46725f365f5fce7597cd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.0625, 13.0, 20.5, 14.3125, 13.375, 12.5625], "runtime_probabilities": [0.0015918540302664042, 0.0005501300329342484, 0.9946584105491638, 0.002043980872258544, 0.0008004344999790192, 0.0003551906265784055], "decide_seconds": 0.16904618400076288, "probabilities_f64_minus_runtime_max_abs": 6.468422464322998e-10}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.002315333763131719, "sagan": 0.0006633542294695925, "socrates": 0.0005854080528233063, "ibn_sina": 0.0010274231293296043, "darwin": 0.9943147945971103, "hypatia": 0.0010936862281354111}, "confidence": 0.9757460970506534, "choice": "darwin"}}, "usage": {"input_tokens": 269, "output_tokens": 1}, "latency_ms": 171.93315099939355, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 269, "compiled_prompt_sha256": "23be69ba5cb1f451624fd3b7c40ae82e0b4db49b8ca665f6aeb2c3876ffe12a3", "compiled_ids_sha256": "3b3b7a4bfea2702a58446afebffef4fbca608d940bd17511298ea5204d64d231", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.1875, 12.9375, 12.8125, 13.375, 20.25, 13.4375], "runtime_probabilities": [0.0023153338115662336, 0.0006633542943745852, 0.0005854080663993955, 0.0010274231899529696, 0.9943148493766785, 0.0010936863254755735], "decide_seconds": 0.1685165679991769, "probabilities_f64_minus_runtime_max_abs": 5.4779568214868846e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0016937827221539725, "ibn_sina": 0.0008516874389205978, "hypatia": 0.0005853556456510678, "einstein": 0.0014041951965365058, "sagan": 0.0012391979115676335, "darwin": 0.9942257810851703}, "confidence": 0.9751866185444787, "choice": "darwin"}}, "usage": {"input_tokens": 269, "output_tokens": 1}, "latency_ms": 171.4448849998007, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 269, "compiled_prompt_sha256": "07e98ff80964b0a6def8b7573c24437abd99e5a530b4fe072fb31b28b2c86dd1", "compiled_ids_sha256": "7f03f77ae1e8fc1b0036d763bbb74bb88191536eb1211803850576b5e3fc0582", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.125, 13.4375, 13.0625, 13.9375, 13.8125, 20.5], "runtime_probabilities": [0.0016937828622758389, 0.0008516874513588846, 0.0005853556212969124, 0.0014041952090337873, 0.0012391980271786451, 0.994225800037384], "decide_seconds": 0.16811290199984796, "probabilities_f64_minus_runtime_max_abs": 1.8952213687306596e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:48Z"} {"id": "c-darwin-01-starter-01", "uid": "c014", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "prompt_sha256_by_order": ["9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564"], "request_sha256": "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "compiled_prompt_sha256_by_order": ["5d90f70b3e510382c9f8fe500bb3610f93699a121cbbf98eabe2725ae32a0559", "c50296a587f94f3042fa25945e6694ba84f7209fa29ab299e32c96f7a7792ba9", "21abcebe34b503bfd4819717c520c0bbfacba4a5ed4c68caa6c68a8c6c46a37e", "152457b9222c9769845f06b7dec35e359f8481e3af967922eb5385ecdff51e5f", "731204adf744e802b7b46e492bcbff22e53b2d27a7384a2a31ba5780b3312552", "3697ea310d117c069d18786e9e9dd55893aa44fbfba134fe57fd4ff3a64ebacb"], "compiled_tokens_by_order": [294, 294, 294, 294, 294, 294], "choice": "sagan", "probs": {"darwin": 0.2707938921248683, "einstein": 0.06846741823124583, "hypatia": 0.09961950570753639, "ibn_sina": 0.06846741823124583, "sagan": 0.34770624017212937, "socrates": 0.14494552553297435}, "confidence": 0.10815786391479243, "seconds": 0.17456503566609172, "seconds_all_orders": 1.0473902139965503, "seconds_by_order": [0.1746018899975752, 0.1744886710002902, 0.17447789200014086, 0.17513841599793523, 0.17437067200080492, 0.1743126729998039], "argmax_stable": true, "spread": {"darwin": 0.2044605448202832, "einstein": 0.1848691772314557, "hypatia": 0.06759532707474161, "ibn_sina": 0.062150379759902646, "sagan": 0.1631942276036098, "socrates": 0.08894182207385784}, "spread_2dp": {"darwin": 0.2, "einstein": 0.18, "hypatia": 0.07, "ibn_sina": 0.07, "sagan": 0.15999999999999998, "socrates": 0.08000000000000002}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.2707938921248683, "einstein": 0.06846741823124583, "hypatia": 0.09961950570753639, "ibn_sina": 0.06846741823124583, "sagan": 0.34770624017212937, "socrates": 0.14494552553297435}, {"ibn_sina": 0.08705074284885245, "hypatia": 0.03202417863279477, "sagan": 0.3442923667544412, "socrates": 0.1435224113651501, "darwin": 0.1842864240571663, "einstein": 0.20882387634159516}, {"hypatia": 0.07516552988493516, "darwin": 0.06633334730458512, "sagan": 0.49014082446342966, "ibn_sina": 0.05166046282455835, "einstein": 0.23152613160942906, "socrates": 0.08517370391306259}, {"ibn_sina": 0.09643575966104653, "einstein": 0.08510425919926604, "darwin": 0.23133736130583343, "sagan": 0.38141082829256834, "hypatia": 0.09643575966104653, "socrates": 0.10927603188023914}, {"einstein": 0.19830313508418693, "sagan": 0.32694659685981986, "socrates": 0.12027693134569536, "ibn_sina": 0.07295164651731573, "darwin": 0.22470689075894834, "hypatia": 0.05681479943403382}, {"socrates": 0.17411552598692043, "ibn_sina": 0.034285379901143886, "hypatia": 0.06405352239934003, "einstein": 0.25333659546270154, "sagan": 0.3686025714068351, "darwin": 0.10560640484305901}], "prompt_tokens": 294, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.2707938921248683, "einstein": 0.06846741823124583, "hypatia": 0.09961950570753639, "ibn_sina": 0.06846741823124583, "sagan": 0.34770624017212937, "socrates": 0.14494552553297435}, "confidence": 0.10815786391479243, "choice": "sagan"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 174.5765610030503, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 294, "compiled_prompt_sha256": "5d90f70b3e510382c9f8fe500bb3610f93699a121cbbf98eabe2725ae32a0559", "compiled_ids_sha256": "1360cb8f139ab7df74cce524630cc666c0ab3b9f2b63c789ee0517529cc2e1d3", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.625, 17.25, 17.625, 17.25, 18.875, 18.0], "runtime_probabilities": [0.2707939147949219, 0.06846741586923599, 0.09961950778961182, 0.06846741586923599, 0.34770625829696655, 0.14494553208351135], "decide_seconds": 0.1707963410008233, "probabilities_f64_minus_runtime_max_abs": 2.2670053556606007e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.08705074284885245, "hypatia": 0.03202417863279477, "sagan": 0.3442923667544412, "socrates": 0.1435224113651501, "darwin": 0.1842864240571663, "einstein": 0.20882387634159516}, "confidence": 0.10301055156154804, "choice": "sagan"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 174.46702099914546, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 294, "compiled_prompt_sha256": "c50296a587f94f3042fa25945e6694ba84f7209fa29ab299e32c96f7a7792ba9", "compiled_ids_sha256": "ab5288f2086e36d2ce132d8350f718f23d1b8eda8cf283b107932631cf036edc", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 16.5, 18.875, 18.0, 18.25, 18.375], "runtime_probabilities": [0.08705074340105057, 0.03202417865395546, 0.34429237246513367, 0.14352241158485413, 0.184286430478096, 0.20882388949394226], "decide_seconds": 0.17089338899677387, "probabilities_f64_minus_runtime_max_abs": 1.3152347100309214e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.07516552988493516, "darwin": 0.06633334730458512, "sagan": 0.49014082446342966, "ibn_sina": 0.05166046282455835, "einstein": 0.23152613160942906, "socrates": 0.08517370391306259}, "confidence": 0.204357927589637, "choice": "sagan"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 174.45561199929216, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 294, "compiled_prompt_sha256": "21abcebe34b503bfd4819717c520c0bbfacba4a5ed4c68caa6c68a8c6c46a37e", "compiled_ids_sha256": "ab09642479d25cc586ce7d8b4685f7d699f98eaf88b15cdd2c1229750d764a16", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 17.25, 19.25, 17.0, 18.5, 17.5], "runtime_probabilities": [0.0751655325293541, 0.0663333460688591, 0.49014079570770264, 0.051660459488630295, 0.23152613639831543, 0.08517369627952576], "decide_seconds": 0.17070862099717488, "probabilities_f64_minus_runtime_max_abs": 2.875572702176754e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.09643575966104653, "einstein": 0.08510425919926604, "darwin": 0.23133736130583343, "sagan": 0.38141082829256834, "hypatia": 0.09643575966104653, "socrates": 0.10927603188023914}, "confidence": 0.1020018750768169, "choice": "sagan"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 175.11670699968818, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 294, "compiled_prompt_sha256": "152457b9222c9769845f06b7dec35e359f8481e3af967922eb5385ecdff51e5f", "compiled_ids_sha256": "97b3d68f937b9608483f5e205e4a89e2c99c250e36d783d5e81e3bb2e66acffc", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 17.5, 18.5, 19.0, 17.625, 17.75], "runtime_probabilities": [0.09643576294183731, 0.08510426431894302, 0.23133736848831177, 0.3814108371734619, 0.09643576294183731, 0.10927603393793106], "decide_seconds": 0.17156321500078775, "probabilities_f64_minus_runtime_max_abs": 8.880893576090898e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.19830313508418693, "sagan": 0.32694659685981986, "socrates": 0.12027693134569536, "ibn_sina": 0.07295164651731573, "darwin": 0.22470689075894834, "hypatia": 0.05681479943403382}, "confidence": 0.08999761771177439, "choice": "sagan"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 174.3479920005484, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 294, "compiled_prompt_sha256": "731204adf744e802b7b46e492bcbff22e53b2d27a7384a2a31ba5780b3312552", "compiled_ids_sha256": "d2281dcb59f7cde98cfaf6d15ff30e234f166bfc18a9ec2df55ab2c3157cfbe5", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.25, 18.75, 17.75, 17.25, 18.375, 17.0], "runtime_probabilities": [0.19830314815044403, 0.3269466161727905, 0.12027695029973984, 0.07295165210962296, 0.22470691800117493, 0.05681479722261429], "decide_seconds": 0.17069152099793428, "probabilities_f64_minus_runtime_max_abs": 2.7242226585677543e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.17411552598692043, "ibn_sina": 0.034285379901143886, "hypatia": 0.06405352239934003, "einstein": 0.25333659546270154, "sagan": 0.3686025714068351, "darwin": 0.10560640484305901}, "confidence": 0.1354009287863096, "choice": "sagan"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 174.29090299992822, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 294, "compiled_prompt_sha256": "3697ea310d117c069d18786e9e9dd55893aa44fbfba134fe57fd4ff3a64ebacb", "compiled_ids_sha256": "2f65e8c958cef8239d6109a8964eda844c5b24bd8ffb8d3f789ab6f2bfa614d4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 16.5, 17.125, 18.5, 18.875, 17.625], "runtime_probabilities": [0.17411553859710693, 0.03428537771105766, 0.06405352056026459, 0.2533366084098816, 0.36860257387161255, 0.10560640692710876], "decide_seconds": 0.1706991309984005, "probabilities_f64_minus_runtime_max_abs": 1.2947180050293383e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:49Z"} {"id": "c-darwin-01-starter-02", "uid": "c015", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "prompt_sha256_by_order": ["30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb"], "request_sha256": "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "compiled_prompt_sha256_by_order": ["cc1df5806cd5c60ea952970a2eda3feb0945e82bb626caee1c5af7534e8df1ee", "52148491ffe991ea9c95419de43cd803e23eb70d293c15b71d16ccbdda347391", "fab59394c58f9dd65f213d2fe89491615f1c92b5ef3dbf6bbfec59942cb9f7ba", "3aeab417b11636c4046dba4e7af465fe2b636584ded218286c084cf35db3f09d", "f9ff1c68587c98be1aafb24549421ccc60d759a186e65e7d9c32de9712e17b3a", "ec8e25a98b9fbce7af2926d67b2dc722125fcc540a8b24ac2e530d7389875c2e"], "compiled_tokens_by_order": [273, 273, 273, 273, 273, 273], "choice": "darwin", "probs": {"darwin": 0.7640990497858285, "einstein": 0.026146031687403076, "hypatia": 0.0230737919790119, "ibn_sina": 0.13278050506802144, "sagan": 0.03804225163150692, "socrates": 0.015858369848228006}, "confidence": 0.5278368457213193, "seconds": 0.17218939450018902, "seconds_all_orders": 1.0331363670011342, "seconds_by_order": [0.17192408199844067, 0.17168006299834815, 0.1725655460031703, 0.17230242900041048, 0.17208010100148385, 0.17258414599928074], "argmax_stable": true, "spread": {"darwin": 0.18397870468694766, "einstein": 0.0762700244610604, "hypatia": 0.02708347180978451, "ibn_sina": 0.12298292651855246, "sagan": 0.03391778579023109, "socrates": 0.013958023680291165}, "spread_2dp": {"darwin": 0.18999999999999995, "einstein": 0.08, "hypatia": 0.03, "ibn_sina": 0.12000000000000001, "sagan": 0.04, "socrates": 0.01}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.7640990497858285, "einstein": 0.026146031687403076, "hypatia": 0.0230737919790119, "ibn_sina": 0.13278050506802144, "sagan": 0.03804225163150692, "socrates": 0.015858369848228006}, {"ibn_sina": 0.017246723945853505, "hypatia": 0.004360654711425643, "sagan": 0.011853488462252648, "socrates": 0.007653187358683441, "darwin": 0.9416392215759313, "einstein": 0.017246723945853505}, {"hypatia": 0.006678506838355313, "darwin": 0.9311258406888111, "sagan": 0.01034387374087996, "ibn_sina": 0.021897980881282284, "einstein": 0.021897980881282284, "socrates": 0.008055816969389056}, {"ibn_sina": 0.009797578549468987, "einstein": 0.009203973273312273, "darwin": 0.9388307972569043, "sagan": 0.01719528586017987, "hypatia": 0.015174786510665427, "socrates": 0.009797578549468987}, {"einstein": 0.011934537630190723, "sagan": 0.0041244658412758345, "socrates": 0.011211460548440149, "ibn_sina": 0.019676726046866544, "darwin": 0.9480777544727762, "hypatia": 0.004975055460450465}, {"socrates": 0.021611211038974606, "ibn_sina": 0.019071826802997106, "hypatia": 0.031444126521210154, "einstein": 0.08547399773437268, "sagan": 0.031444126521210154, "darwin": 0.8109547113812352}], "prompt_tokens": 273, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.7640990497858285, "einstein": 0.026146031687403076, "hypatia": 0.0230737919790119, "ibn_sina": 0.13278050506802144, "sagan": 0.03804225163150692, "socrates": 0.015858369848228006}, "confidence": 0.5278368457213193, "choice": "darwin"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 171.90049199780333, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 273, "compiled_prompt_sha256": "cc1df5806cd5c60ea952970a2eda3feb0945e82bb626caee1c5af7534e8df1ee", "compiled_ids_sha256": "f791a198d080ca74c8294b34a536763716401dc9291fa42c4a1acd27960ce18f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.25, 16.875, 16.75, 18.5, 17.25, 16.375], "runtime_probabilities": [0.7640990614891052, 0.02614603191614151, 0.023073792457580566, 0.13278049230575562, 0.03804225102066994, 0.015858368948101997], "decide_seconds": 0.16847236800094834, "probabilities_f64_minus_runtime_max_abs": 1.2762265827470642e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.017246723945853505, "hypatia": 0.004360654711425643, "sagan": 0.011853488462252648, "socrates": 0.007653187358683441, "darwin": 0.9416392215759313, "einstein": 0.017246723945853505}, "confidence": 0.8268543507773649, "choice": "darwin"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 171.6591230033373, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 273, "compiled_prompt_sha256": "52148491ffe991ea9c95419de43cd803e23eb70d293c15b71d16ccbdda347391", "compiled_ids_sha256": "aed46d4b36b34ae538fa0df4c24ada9c8eedf149a9a24196484367a9c22f3e5e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 14.875, 15.875, 15.4375, 20.25, 16.25], "runtime_probabilities": [0.017246725037693977, 0.004360654857009649, 0.01185348816215992, 0.007653187960386276, 0.9416391849517822, 0.017246725037693977], "decide_seconds": 0.1683519799989881, "probabilities_f64_minus_runtime_max_abs": 3.6624149046815546e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.006678506838355313, "darwin": 0.9311258406888111, "sagan": 0.01034387374087996, "ibn_sina": 0.021897980881282284, "einstein": 0.021897980881282284, "socrates": 0.008055816969389056}, "confidence": 0.802772974696972, "choice": "darwin"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 172.54433600101038, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 273, "compiled_prompt_sha256": "fab59394c58f9dd65f213d2fe89491615f1c92b5ef3dbf6bbfec59942cb9f7ba", "compiled_ids_sha256": "d0e4e9e90aaca1119fab0b84cb552a96f5a742ea58ecad775b431dd1e725d610", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.6875, 20.625, 16.125, 16.875, 16.875, 15.875], "runtime_probabilities": [0.006678505800664425, 0.9311257600784302, 0.010343873873353004, 0.021897979080677032, 0.021897979080677032, 0.008055816404521465], "decide_seconds": 0.1690667740003846, "probabilities_f64_minus_runtime_max_abs": 8.061038092588291e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.009797578549468987, "einstein": 0.009203973273312273, "darwin": 0.9388307972569043, "sagan": 0.01719528586017987, "hypatia": 0.015174786510665427, "socrates": 0.009797578549468987}, "confidence": 0.8177945004993887, "choice": "darwin"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 172.28168899964658, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 273, "compiled_prompt_sha256": "3aeab417b11636c4046dba4e7af465fe2b636584ded218286c084cf35db3f09d", "compiled_ids_sha256": "4ed0122584c876807f82edaead790265a969218b57f14fb5de3c5144ecf5106a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 15.875, 20.5, 16.5, 16.375, 15.9375], "runtime_probabilities": [0.009797578677535057, 0.009203973226249218, 0.9388307929039001, 0.017195286229252815, 0.015174786560237408, 0.009797578677535057], "decide_seconds": 0.1689741039990622, "probabilities_f64_minus_runtime_max_abs": 4.35300417933604e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.011934537630190723, "sagan": 0.0041244658412758345, "socrates": 0.011211460548440149, "ibn_sina": 0.019676726046866544, "darwin": 0.9480777544727762, "hypatia": 0.004975055460450465}, "confidence": 0.8436864091741947, "choice": "darwin"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 172.05869099780102, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 273, "compiled_prompt_sha256": "f9ff1c68587c98be1aafb24549421ccc60d759a186e65e7d9c32de9712e17b3a", "compiled_ids_sha256": "146c264016969c6ab78aae7142bd6a241931b9b9f80907562acb25accf5a4a89", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 14.9375, 15.9375, 16.5, 20.375, 15.125], "runtime_probabilities": [0.011934537440538406, 0.004124465864151716, 0.011211459524929523, 0.019676724448800087, 0.9480777382850647, 0.00497505534440279], "decide_seconds": 0.16868931600038195, "probabilities_f64_minus_runtime_max_abs": 1.6187711504578317e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.021611211038974606, "ibn_sina": 0.019071826802997106, "hypatia": 0.031444126521210154, "einstein": 0.08547399773437268, "sagan": 0.031444126521210154, "darwin": 0.8109547113812352}, "confidence": 0.5780089960944341, "choice": "darwin"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 172.56286699921475, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 273, "compiled_prompt_sha256": "ec8e25a98b9fbce7af2926d67b2dc722125fcc540a8b24ac2e530d7389875c2e", "compiled_ids_sha256": "12537b2200ce45c8f300bfacf497548160c14918c5fd253f2a92610b46f0a5cf", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 16.375, 16.875, 17.875, 16.875, 20.125], "runtime_probabilities": [0.021611209958791733, 0.019071826711297035, 0.03144412487745285, 0.08547399193048477, 0.03144412487745285, 0.8109546899795532], "decide_seconds": 0.16904811300264555, "probabilities_f64_minus_runtime_max_abs": 2.1401681982169407e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:50Z"} {"id": "c-darwin-01-starter-03", "uid": "c016", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "prompt_sha256_by_order": ["e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8"], "request_sha256": "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "compiled_prompt_sha256_by_order": ["48c20dba52dff386d2809ed467785e33fb1eac6e82e19292cd9497781a43ff1b", "d621d8ab1a56a4ed90e1e7646fe59f2f49b01db39f68788e02a52704a33c967a", "46d735edbb9564b6392a1fd9864b5af96d1e82972852bcd9684c302d3b94af77", "7db96ad157113e2bb7ea1485d276629de6e7b9334aceddb0d88b230da827d363", "a8473c49657743c60018490037305ac5afd58b210cdd3e1526fc0e0c5cfd37e6", "5047cc060e275b6fe6eb8ee0793f1860ce1a695c773545ccf05ee1db882b088e"], "compiled_tokens_by_order": [396, 396, 396, 396, 396, 396], "choice": "darwin", "probs": {"darwin": 0.9104402537997479, "einstein": 0.004777551983169547, "hypatia": 0.027492913427103957, "ibn_sina": 0.027492913427103957, "sagan": 0.008384864856931623, "socrates": 0.0214115025059429}, "confidence": 0.7594784686253963, "seconds": 0.21152443300040127, "seconds_all_orders": 1.2691465980024077, "seconds_by_order": [0.21158257299975958, 0.211573672000668, 0.2117868109999108, 0.21141057400018326, 0.21100410700091743, 0.2117888610009686], "argmax_stable": true, "spread": {"darwin": 0.05883863880268425, "einstein": 0.01791686234668473, "hypatia": 0.03250500713838192, "ibn_sina": 0.03224122094511477, "sagan": 0.015376260379025229, "socrates": 0.036571987676422044}, "spread_2dp": {"darwin": 0.06000000000000005, "einstein": 0.02, "hypatia": 0.030000000000000002, "ibn_sina": 0.030000000000000002, "sagan": 0.02, "socrates": 0.039999999999999994}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.9104402537997479, "einstein": 0.004777551983169547, "hypatia": 0.027492913427103957, "ibn_sina": 0.027492913427103957, "sagan": 0.008384864856931623, "socrates": 0.0214115025059429}, {"ibn_sina": 0.03968040119035179, "hypatia": 0.015539068374934512, "sagan": 0.014597603815365248, "socrates": 0.01874369432166693, "darwin": 0.9031217683614597, "einstein": 0.008317463936221712}, {"hypatia": 0.030706281184373626, "darwin": 0.8973690751985298, "sagan": 0.012800282391334171, "ibn_sina": 0.023914075831600927, "einstein": 0.011296209562560545, "socrates": 0.023914075831600927}, {"ibn_sina": 0.05493563527496905, "einstein": 0.010162070717960863, "darwin": 0.8593379199303633, "sagan": 0.02020969080535379, "hypatia": 0.029404926613878165, "socrates": 0.02594975665747482}, {"einstein": 0.017958372644289438, "sagan": 0.004833430426328561, "socrates": 0.05531568199808898, "ibn_sina": 0.023059006917617482, "darwin": 0.8652828509189204, "hypatia": 0.03355065709475515}, {"socrates": 0.04804407551331643, "ibn_sina": 0.02269441432985428, "hypatia": 0.04804407551331643, "einstein": 0.02269441432985428, "sagan": 0.006921405316594896, "darwin": 0.8516016149970637}], "prompt_tokens": 396, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.9104402537997479, "einstein": 0.004777551983169547, "hypatia": 0.027492913427103957, "ibn_sina": 0.027492913427103957, "sagan": 0.008384864856931623, "socrates": 0.0214115025059429}, "confidence": 0.7594784686253963, "choice": "darwin"}}, "usage": {"input_tokens": 396, "output_tokens": 1}, "latency_ms": 211.554082998191, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 396, "compiled_prompt_sha256": "48c20dba52dff386d2809ed467785e33fb1eac6e82e19292cd9497781a43ff1b", "compiled_ids_sha256": "ec6e841b4247aaa12db66e4d4109704bc2a69c4a9bbb1d5a4325d9879d6dca16", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.5, 15.25, 17.0, 17.0, 15.8125, 16.75], "runtime_probabilities": [0.9104402661323547, 0.004777552094310522, 0.027492914348840714, 0.027492914348840714, 0.008384864777326584, 0.021411502733826637], "decide_seconds": 0.20669782999902964, "probabilities_f64_minus_runtime_max_abs": 1.2332606824649872e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.03968040119035179, "hypatia": 0.015539068374934512, "sagan": 0.014597603815365248, "socrates": 0.01874369432166693, "darwin": 0.9031217683614597, "einstein": 0.008317463936221712}, "confidence": 0.7427882166571803, "choice": "darwin"}}, "usage": {"input_tokens": 396, "output_tokens": 1}, "latency_ms": 211.54883399867686, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 396, "compiled_prompt_sha256": "d621d8ab1a56a4ed90e1e7646fe59f2f49b01db39f68788e02a52704a33c967a", "compiled_ids_sha256": "77760dc08b2e87ff16d3e3c31f88c38202a36be0abcd975e12fcfc15ffb768ef", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 15.9375, 15.875, 16.125, 20.0, 15.3125], "runtime_probabilities": [0.03968040272593498, 0.015539068728685379, 0.014597604051232338, 0.018743693828582764, 0.903121829032898, 0.008317464031279087], "decide_seconds": 0.2068972089982708, "probabilities_f64_minus_runtime_max_abs": 6.06714382067608e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.030706281184373626, "darwin": 0.8973690751985298, "sagan": 0.012800282391334171, "ibn_sina": 0.023914075831600927, "einstein": 0.011296209562560545, "socrates": 0.023914075831600927}, "confidence": 0.7270164212033379, "choice": "darwin"}}, "usage": {"input_tokens": 396, "output_tokens": 1}, "latency_ms": 211.76176099834265, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 396, "compiled_prompt_sha256": "46d735edbb9564b6392a1fd9864b5af96d1e82972852bcd9684c302d3b94af77", "compiled_ids_sha256": "107564a7bfa8934860b365ef9a20b4770878cda01b401cc010871b64d7b4e077", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 20.5, 16.25, 16.875, 16.125, 16.875], "runtime_probabilities": [0.03070628084242344, 0.8973690867424011, 0.01280028186738491, 0.023914076387882233, 0.011296209879219532, 0.023914076387882233], "decide_seconds": 0.206949749001069, "probabilities_f64_minus_runtime_max_abs": 1.1543871303310027e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.05493563527496905, "einstein": 0.010162070717960863, "darwin": 0.8593379199303633, "sagan": 0.02020969080535379, "hypatia": 0.029404926613878165, "socrates": 0.02594975665747482}, "confidence": 0.6575364275875475, "choice": "darwin"}}, "usage": {"input_tokens": 396, "output_tokens": 1}, "latency_ms": 211.38518400039175, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 396, "compiled_prompt_sha256": "7db96ad157113e2bb7ea1485d276629de6e7b9334aceddb0d88b230da827d363", "compiled_ids_sha256": "1362790190d182c95fc794b8ae3ebf66ae9e00a934976662d5b19c9c7c4c9410", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 15.8125, 20.25, 16.5, 16.875, 16.75], "runtime_probabilities": [0.05493563786149025, 0.010162070393562317, 0.8593379855155945, 0.020209692418575287, 0.029404927045106888, 0.025949757546186447], "decide_seconds": 0.20667906100061373, "probabilities_f64_minus_runtime_max_abs": 6.558523113486814e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.017958372644289438, "sagan": 0.004833430426328561, "socrates": 0.05531568199808898, "ibn_sina": 0.023059006917617482, "darwin": 0.8652828509189204, "hypatia": 0.03355065709475515}, "confidence": 0.6740033382195609, "choice": "darwin"}}, "usage": {"input_tokens": 396, "output_tokens": 1}, "latency_ms": 210.97860699956072, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 396, "compiled_prompt_sha256": "a8473c49657743c60018490037305ac5afd58b210cdd3e1526fc0e0c5cfd37e6", "compiled_ids_sha256": "1afcadb575843ea5e49305250c2aeb0f443a6a59cd584fe0bab9d9431f0fc45b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 14.8125, 17.25, 16.375, 20.0, 16.75], "runtime_probabilities": [0.017958370968699455, 0.004833430517464876, 0.05531567707657814, 0.023059004917740822, 0.8652828335762024, 0.033550653606653214], "decide_seconds": 0.20635962399683194, "probabilities_f64_minus_runtime_max_abs": 1.734271803943699e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.04804407551331643, "ibn_sina": 0.02269441432985428, "hypatia": 0.04804407551331643, "einstein": 0.02269441432985428, "sagan": 0.006921405316594896, "darwin": 0.8516016149970637}, "confidence": 0.6457484045052104, "choice": "darwin"}}, "usage": {"input_tokens": 396, "output_tokens": 1}, "latency_ms": 211.76447099787765, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 396, "compiled_prompt_sha256": "5047cc060e275b6fe6eb8ee0793f1860ce1a695c773545ccf05ee1db882b088e", "compiled_ids_sha256": "326c215ba64135981dbd53c35d2ce479342b5155494c4d771b2d0043d068ea5e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.25, 16.5, 17.25, 16.5, 15.3125, 20.125], "runtime_probabilities": [0.048044078052043915, 0.022694414481520653, 0.048044078052043915, 0.022694414481520653, 0.006921405903995037, 0.8516016006469727], "decide_seconds": 0.20716089800043846, "probabilities_f64_minus_runtime_max_abs": 1.4350091004367016e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:51Z"} {"id": "c-darwin-01-starter-03", "uid": "c017", "task": "kperm", "label": "darwin", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "prompt_sha256_by_order": ["d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169"], "request_sha256": "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "compiled_prompt_sha256_by_order": ["292b2a1853810df91bacbfc1e9d9b69a191c104ded5e174b75e99205cda89199", "13e350e2ee4c4c6afc6af2ffd85fbda214625bb4dc1433ad0c22fae70f15c73a", "391e77e8f3485628fc7e8f6848ca2671886f500c6b41a0cf8eb9eb1fd066b019", "354e5e656727557de97828ddafe9b8576c3a48546f60c17c4401fc471d05e288", "12ea3388b537ba1ca5581a269d64e136849aa4fdf8c9e47fa9538a59c8c34e7c", "5dfff57f55994310a2473a4248ea85d070dd2fa08d515240a3561ad357d461b0"], "compiled_tokens_by_order": [383, 383, 383, 383, 383, 383], "choice": "darwin", "probs": {"darwin": 0.9686427650615624, "einstein": 0.00255587819389861, "hypatia": 0.007395678775662063, "ibn_sina": 0.015656652090938594, "sagan": 0.003493471288439377, "socrates": 0.0022555545894990332}, "confidence": 0.8989833021344859, "seconds": 0.19569864799935507, "seconds_all_orders": 1.1741918879961304, "seconds_by_order": [0.19573073599894997, 0.19615250299830223, 0.19543089799844893, 0.19539727899973514, 0.1959057739986747, 0.19557469800201943], "argmax_stable": true, "spread": {"darwin": 0.017593221017406457, "einstein": 0.005173791706404563, "hypatia": 0.0088787624920156, "ibn_sina": 0.005480457432167971, "sagan": 0.00410442832774071, "socrates": 0.007669538025096442}, "spread_2dp": {"darwin": 0.020000000000000018, "einstein": 0.01, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.01, "socrates": 0.01}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.9686427650615624, "einstein": 0.00255587819389861, "hypatia": 0.007395678775662063, "ibn_sina": 0.015656652090938594, "sagan": 0.003493471288439377, "socrates": 0.0022555545894990332}, {"ibn_sina": 0.01765122194799021, "hypatia": 0.006493521666219737, "sagan": 0.0034757316831797898, "socrates": 0.004462927822789714, "darwin": 0.9637240641846981, "einstein": 0.004192532695122363}, {"hypatia": 0.008340331996761515, "darwin": 0.9640113065648902, "sagan": 0.004464258018909211, "ibn_sina": 0.013750882767761584, "einstein": 0.005732220762931491, "socrates": 0.0037009998887460845}, {"ibn_sina": 0.013726846066896928, "einstein": 0.0034706902112281927, "darwin": 0.9623262037393195, "sagan": 0.006091251018082563, "hypatia": 0.01069047846599996, "socrates": 0.003694530498472867}, {"einstein": 0.006934692584960079, "sagan": 0.0019868226903418536, "socrates": 0.004206103668460144, "ibn_sina": 0.012170764515822238, "darwin": 0.9668435803652741, "hypatia": 0.007858036175141504}, {"socrates": 0.009925092614595475, "ibn_sina": 0.013565993155292931, "hypatia": 0.015372284158235338, "einstein": 0.007729669900303173, "sagan": 0.0023574161274169995, "darwin": 0.951049544044156}], "prompt_tokens": 383, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.9686427650615624, "einstein": 0.00255587819389861, "hypatia": 0.007395678775662063, "ibn_sina": 0.015656652090938594, "sagan": 0.003493471288439377, "socrates": 0.0022555545894990332}, "confidence": 0.8989833021344859, "choice": "darwin"}}, "usage": {"input_tokens": 383, "output_tokens": 1}, "latency_ms": 195.69440700070118, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 383, "compiled_prompt_sha256": "292b2a1853810df91bacbfc1e9d9b69a191c104ded5e174b75e99205cda89199", "compiled_ids_sha256": "80f60e0dbaddd871d7e686f5baae1c6e910fcda7924288511f08aec576cdfe5e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.75, 14.8125, 15.875, 16.625, 15.125, 14.6875], "runtime_probabilities": [0.9686427712440491, 0.002555878134444356, 0.007395678199827671, 0.01565665192902088, 0.0034934713039547205, 0.0022555545438081026], "decide_seconds": 0.19106494200241286, "probabilities_f64_minus_runtime_max_abs": 6.182486633754536e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.01765122194799021, "hypatia": 0.006493521666219737, "sagan": 0.0034757316831797898, "socrates": 0.004462927822789714, "darwin": 0.9637240641846981, "einstein": 0.004192532695122363}, "confidence": 0.8848289127080246, "choice": "darwin"}}, "usage": {"input_tokens": 383, "output_tokens": 1}, "latency_ms": 196.12444300219067, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 383, "compiled_prompt_sha256": "13e350e2ee4c4c6afc6af2ffd85fbda214625bb4dc1433ad0c22fae70f15c73a", "compiled_ids_sha256": "8f3957dd490479edf1ddf6cd43137e37f5d81c933d3083c3971e4b77a8d53bca", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 15.25, 14.625, 14.875, 20.25, 14.8125], "runtime_probabilities": [0.01765122264623642, 0.0064935218542814255, 0.0034757317043840885, 0.004462928045541048, 0.9637240767478943, 0.00419253297150135], "decide_seconds": 0.1914866390034149, "probabilities_f64_minus_runtime_max_abs": 1.2563196150061628e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.008340331996761515, "darwin": 0.9640113065648902, "sagan": 0.004464258018909211, "ibn_sina": 0.013750882767761584, "einstein": 0.005732220762931491, "socrates": 0.0037009998887460845}, "confidence": 0.8835391032598051, "choice": "darwin"}}, "usage": {"input_tokens": 383, "output_tokens": 1}, "latency_ms": 195.40387799861492, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 383, "compiled_prompt_sha256": "391e77e8f3485628fc7e8f6848ca2671886f500c6b41a0cf8eb9eb1fd066b019", "compiled_ids_sha256": "3b2209be9db2c78c261e60daf8d52264fe6f390eb99a928b67ec300f1e905340", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 20.875, 15.5, 16.625, 15.75, 15.3125], "runtime_probabilities": [0.00834033265709877, 0.9640113711357117, 0.004464258439838886, 0.01375088281929493, 0.005732220597565174, 0.003701000241562724], "decide_seconds": 0.19089539400010835, "probabilities_f64_minus_runtime_max_abs": 6.457082146749116e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.013726846066896928, "einstein": 0.0034706902112281927, "darwin": 0.9623262037393195, "sagan": 0.006091251018082563, "hypatia": 0.01069047846599996, "socrates": 0.003694530498472867}, "confidence": 0.8795829094977422, "choice": "darwin"}}, "usage": {"input_tokens": 383, "output_tokens": 1}, "latency_ms": 195.3703389990551, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 383, "compiled_prompt_sha256": "354e5e656727557de97828ddafe9b8576c3a48546f60c17c4401fc471d05e288", "compiled_ids_sha256": "d044de91ccf34555b678168d7bc293e186fa5431d2b77c7f70010d28e63c4a60", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 14.875, 20.5, 15.4375, 16.0, 14.9375], "runtime_probabilities": [0.013726845383644104, 0.0034706902224570513, 0.962326169013977, 0.0060912505723536015, 0.0106904786080122, 0.003694530576467514], "decide_seconds": 0.19091838399981498, "probabilities_f64_minus_runtime_max_abs": 3.4725342490382616e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.006934692584960079, "sagan": 0.0019868226903418536, "socrates": 0.004206103668460144, "ibn_sina": 0.012170764515822238, "darwin": 0.9668435803652741, "hypatia": 0.007858036175141504}, "confidence": 0.8916223688690795, "choice": "darwin"}}, "usage": {"input_tokens": 383, "output_tokens": 1}, "latency_ms": 195.87927400061744, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 383, "compiled_prompt_sha256": "12ea3388b537ba1ca5581a269d64e136849aa4fdf8c9e47fa9538a59c8c34e7c", "compiled_ids_sha256": "a0b56cf8462290e5e32f8970810e85c61cee48ba7cfdaa90b873eb42285b83b4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.4375, 14.1875, 14.9375, 16.0, 20.375, 15.5625], "runtime_probabilities": [0.006934692617505789, 0.001986822811886668, 0.00420610373839736, 0.012170765548944473, 0.9668436050415039, 0.007858037017285824], "decide_seconds": 0.19137210000189953, "probabilities_f64_minus_runtime_max_abs": 2.4676229837261587e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.009925092614595475, "ibn_sina": 0.013565993155292931, "hypatia": 0.015372284158235338, "einstein": 0.007729669900303173, "sagan": 0.0023574161274169995, "darwin": 0.951049544044156}, "confidence": 0.8504921911204057, "choice": "darwin"}}, "usage": {"input_tokens": 383, "output_tokens": 1}, "latency_ms": 195.54845799939358, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 383, "compiled_prompt_sha256": "5dfff57f55994310a2473a4248ea85d070dd2fa08d515240a3561ad357d461b0", "compiled_ids_sha256": "6e87572a499b5ef32fba05091c0ae7c63d9cd0c298c0a8266754befaf9c58f1c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 16.25, 16.375, 15.6875, 14.5, 20.5], "runtime_probabilities": [0.00992509350180626, 0.013565992936491966, 0.01537228375673294, 0.007729670498520136, 0.002357416320592165, 0.9510495662689209], "decide_seconds": 0.19110746200021822, "probabilities_f64_minus_runtime_max_abs": 2.222476491731129e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:52Z"} {"id": "c-einstein-01-dessert-01", "uid": "c018", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "prompt_sha256_by_order": ["f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3"], "request_sha256": "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "compiled_prompt_sha256_by_order": ["511f397ccbe1d23a7056a582c58d20ec2dc37e7b408ccc37042cd3604b440fdb", "16420fcd2d02a093c6ffb478645762f51ec9e0a475f29d62048c3b5eaea2a2ab", "399555e241d43dd37e0db5aa8d6ff1a45f2de6766a64bb3c013770f4ae212881", "8ceabf110cc29e0400fc03f2bb9ab29ab7a8a40842ad988676de7d57a6b7af10", "8f6060573b36acad3b9b762461c3cf0d95ab019b9cb99e5b823a2cd0ff0802e8", "ef3fabe0a28442c41a3c1edd976bdc9de84545f815124210cbb69daad8e48260"], "compiled_tokens_by_order": [373, 373, 373, 373, 373, 373], "choice": "sagan", "probs": {"darwin": 0.004673603521077249, "einstein": 0.041655374755950325, "hypatia": 0.00283468382687427, "ibn_sina": 0.0014253685417158948, "sagan": 0.9480719595269709, "socrates": 0.0013390098274113063}, "confidence": 0.86445994053331, "seconds": 0.19477208183282832, "seconds_all_orders": 1.1686324909969699, "seconds_by_order": [0.19504954099829774, 0.1949167029997625, 0.19491257300251164, 0.1940327989977959, 0.19467258399890852, 0.19504829099969356], "argmax_stable": false, "spread": {"darwin": 0.013778438977126443, "einstein": 0.47192333782796875, "hypatia": 0.003261633570266494, "ibn_sina": 0.0036426329040966524, "sagan": 0.4732819516733525, "socrates": 0.009389949317566915}, "spread_2dp": {"darwin": 0.02, "einstein": 0.47, "hypatia": 0.0, "ibn_sina": 0.01, "sagan": 0.48, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.004673603521077249, "einstein": 0.041655374755950325, "hypatia": 0.00283468382687427, "ibn_sina": 0.0014253685417158948, "sagan": 0.9480719595269709, "socrates": 0.0013390098274113063}, {"ibn_sina": 0.0021763567750071973, "hypatia": 0.0009657524646861775, "sagan": 0.934631444575968, "socrates": 0.002794497414889805, "darwin": 0.006703650807634077, "einstein": 0.05272829796181476}, {"hypatia": 0.0032154434761551653, "darwin": 0.0018321044666063083, "sagan": 0.89155335838855, "ibn_sina": 0.003422821763330754, "einstein": 0.0939690326296791, "socrates": 0.0060072392756787005}, {"ibn_sina": 0.0030721415070166054, "einstein": 0.02003287907931365, "darwin": 0.005065105049219025, "sagan": 0.9652381685806349, "hypatia": 0.004199119572445141, "socrates": 0.0023925862113706982}, {"einstein": 0.21549690505383207, "sagan": 0.7521581049094659, "socrates": 0.010728959144978222, "ibn_sina": 0.005068001445812547, "darwin": 0.015610543443732751, "hypatia": 0.0009374860021786471}, {"socrates": 0.007469960583617282, "ibn_sina": 0.0017742711564315635, "hypatia": 0.003528559529994554, "einstein": 0.4919562169072824, "sagan": 0.4919562169072824, "darwin": 0.0033147749153918625}], "prompt_tokens": 373, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.004673603521077249, "einstein": 0.041655374755950325, "hypatia": 0.00283468382687427, "ibn_sina": 0.0014253685417158948, "sagan": 0.9480719595269709, "socrates": 0.0013390098274113063}, "confidence": 0.86445994053331, "choice": "sagan"}}, "usage": {"input_tokens": 373, "output_tokens": 1}, "latency_ms": 195.0211310031591, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 373, "compiled_prompt_sha256": "511f397ccbe1d23a7056a582c58d20ec2dc37e7b408ccc37042cd3604b440fdb", "compiled_ids_sha256": "95907c564f0990963bab66233f57176e924cae32017e3e77c3639ed1e429d350", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.3125, 17.5, 14.8125, 14.125, 20.625, 14.0625], "runtime_probabilities": [0.004673603922128677, 0.04165538027882576, 0.0028346842154860497, 0.0014253687113523483, 0.948072075843811, 0.0013390099629759789], "decide_seconds": 0.1905651859997306, "probabilities_f64_minus_runtime_max_abs": 1.1631684015966925e-07}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0021763567750071973, "hypatia": 0.0009657524646861775, "sagan": 0.934631444575968, "socrates": 0.002794497414889805, "darwin": 0.006703650807634077, "einstein": 0.05272829796181476}, "confidence": 0.8390560400832766, "choice": "sagan"}}, "usage": {"input_tokens": 373, "output_tokens": 1}, "latency_ms": 194.8921130024246, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 373, "compiled_prompt_sha256": "16420fcd2d02a093c6ffb478645762f51ec9e0a475f29d62048c3b5eaea2a2ab", "compiled_ids_sha256": "5a60a66c897a80858b5972b6551068a349f2461dd904e13efc7f8ed7d6b5ae99", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.4375, 13.625, 20.5, 14.6875, 15.5625, 17.625], "runtime_probabilities": [0.0021763567347079515, 0.0009657524642534554, 0.9346314668655396, 0.002794497413560748, 0.006703650578856468, 0.05272830277681351], "decide_seconds": 0.1903283880019444, "probabilities_f64_minus_runtime_max_abs": 2.2289571521838525e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0032154434761551653, "darwin": 0.0018321044666063083, "sagan": 0.89155335838855, "ibn_sina": 0.003422821763330754, "einstein": 0.0939690326296791, "socrates": 0.0060072392756787005}, "confidence": 0.7741220404612636, "choice": "sagan"}}, "usage": {"input_tokens": 373, "output_tokens": 1}, "latency_ms": 194.88840299891308, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 373, "compiled_prompt_sha256": "399555e241d43dd37e0db5aa8d6ff1a45f2de6766a64bb3c013770f4ae212881", "compiled_ids_sha256": "5a4e05941fdf182cba2855a4f92d8814a8b5a12c095150aa66d75e2d96a6935f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.75, 14.1875, 20.375, 14.8125, 18.125, 15.375], "runtime_probabilities": [0.0032154435757547617, 0.0018321045208722353, 0.8915534019470215, 0.003422822104766965, 0.09396903216838837, 0.006007239688187838], "decide_seconds": 0.19055939600002603, "probabilities_f64_minus_runtime_max_abs": 4.355847149639658e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0030721415070166054, "einstein": 0.02003287907931365, "darwin": 0.005065105049219025, "sagan": 0.9652381685806349, "hypatia": 0.004199119572445141, "socrates": 0.0023925862113706982}, "confidence": 0.8914739080972238, "choice": "sagan"}}, "usage": {"input_tokens": 373, "output_tokens": 1}, "latency_ms": 194.00798000060604, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 373, "compiled_prompt_sha256": "8ceabf110cc29e0400fc03f2bb9ab29ab7a8a40842ad988676de7d57a6b7af10", "compiled_ids_sha256": "a9c3f5411aea2f5cecb8a7fc2b034aaabeca50234ed8c268fc16da47b71328ab", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.125, 17.0, 15.625, 20.875, 15.4375, 14.875], "runtime_probabilities": [0.003072141669690609, 0.02003287896513939, 0.005065104924142361, 0.9652381539344788, 0.004199119750410318, 0.0023925863206386566], "decide_seconds": 0.18973398199887015, "probabilities_f64_minus_runtime_max_abs": 1.4646156176745251e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.21549690505383207, "sagan": 0.7521581049094659, "socrates": 0.010728959144978222, "ibn_sina": 0.005068001445812547, "darwin": 0.015610543443732751, "hypatia": 0.0009374860021786471}, "confidence": 0.6138554021789282, "choice": "sagan"}}, "usage": {"input_tokens": 373, "output_tokens": 1}, "latency_ms": 194.64877400241676, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 373, "compiled_prompt_sha256": "8f6060573b36acad3b9b762461c3cf0d95ab019b9cb99e5b823a2cd0ff0802e8", "compiled_ids_sha256": "3dbef42a3452a01e857cdda10efc3e47e726fe41a45b3674dbc090ab674062a4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.75, 20.0, 15.75, 15.0, 16.125, 13.3125], "runtime_probabilities": [0.21549689769744873, 0.7521580457687378, 0.010728958062827587, 0.0050680008716881275, 0.015610541217029095, 0.0009374859509989619], "decide_seconds": 0.19030625899904408, "probabilities_f64_minus_runtime_max_abs": 5.914072809254378e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.007469960583617282, "ibn_sina": 0.0017742711564315635, "hypatia": 0.003528559529994554, "einstein": 0.4919562169072824, "sagan": 0.4919562169072824, "darwin": 0.0033147749153918625}, "confidence": 0.5620938845833494, "choice": "einstein"}}, "usage": {"input_tokens": 373, "output_tokens": 1}, "latency_ms": 195.02428099804092, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 373, "compiled_prompt_sha256": "ef3fabe0a28442c41a3c1edd976bdc9de84545f815124210cbb69daad8e48260", "compiled_ids_sha256": "af18a24b0165ecff509fd261726553ffacba047d434f04cc37b16bc02e0570d3", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.4375, 14.0, 14.6875, 19.625, 19.625, 14.625], "runtime_probabilities": [0.007469960954040289, 0.0017742713680490851, 0.0035285595804452896, 0.49195626378059387, 0.49195626378059387, 0.0033147751819342375], "decide_seconds": 0.19074798499786993, "probabilities_f64_minus_runtime_max_abs": 4.687331145447615e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:53Z"} {"id": "c-einstein-01-dessert-01", "uid": "c019", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "prompt_sha256_by_order": ["b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd"], "request_sha256": "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "compiled_prompt_sha256_by_order": ["482a99aea3518461af305b2fde5179c25aa0874e22f2f00e33c301df075be35a", "e8f61a055e57f14754c74acd7cc3bb1a481e3083be49c4f112106963b4d19de2", "83792c4ef43956e9c0615090fdc2a9d10d767913a586f733583dd6c772ddac9f", "fc6239bffaf1148332bf71061007eaa3073fe882d2532f6871925929decbd17b", "6d488638770e470d6e8b24ecb6a22f3ac30f19d16d04d3d47158a932c3d94110", "b0631ff6dd0c6bb01252176fe7b6fb465b6fddfad854d371cab3da4586843feb"], "compiled_tokens_by_order": [380, 380, 380, 380, 380, 380], "choice": "sagan", "probs": {"darwin": 0.008937086954007608, "einstein": 0.010127046256857434, "hypatia": 0.006538512492807102, "ibn_sina": 0.0021227442136432196, "sagan": 0.9704012948891652, "socrates": 0.001873315193519514}, "confidence": 0.9020262349065657, "seconds": 0.1954278184997141, "seconds_all_orders": 1.1725669109982846, "seconds_by_order": [0.1962700619988027, 0.19495145299879368, 0.19534745900091366, 0.19535876900044968, 0.19510485099817743, 0.19553431700114743], "argmax_stable": true, "spread": {"darwin": 0.017203481785756383, "einstein": 0.17176747093126407, "hypatia": 0.005708959573823888, "ibn_sina": 0.0032305722212701833, "sagan": 0.16604347713619572, "socrates": 0.009644539730884711}, "spread_2dp": {"darwin": 0.02, "einstein": 0.16999999999999998, "hypatia": 0.01, "ibn_sina": 0.0, "sagan": 0.15999999999999992, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.008937086954007608, "einstein": 0.010127046256857434, "hypatia": 0.006538512492807102, "ibn_sina": 0.0021227442136432196, "sagan": 0.9704012948891652, "socrates": 0.001873315193519514}, {"ibn_sina": 0.002254958625417229, "hypatia": 0.0008295529189832139, "sagan": 0.968386829648266, "socrates": 0.002118337588820864, "darwin": 0.010757805938242028, "einstein": 0.015652515280270854}, {"hypatia": 0.004091381176832874, "darwin": 0.0021899585333272718, "sagan": 0.9404726886097834, "ibn_sina": 0.003610631215847928, "einstein": 0.046823378045924104, "socrates": 0.002811962418284425}, {"ibn_sina": 0.0029110188847721987, "einstein": 0.008423312438384236, "darwin": 0.010815747263587354, "sagan": 0.9736025415396025, "hypatia": 0.003511359035447954, "socrates": 0.0007360208382057865}, {"einstein": 0.028217289164177185, "sagan": 0.9344282837208389, "socrates": 0.010380560569090497, "ibn_sina": 0.004606345829667633, "darwin": 0.019393440319083653, "hypatia": 0.0029740803971421725}, {"socrates": 0.003980932933357408, "ibn_sina": 0.0013757736083974494, "hypatia": 0.003980932933357408, "einstein": 0.1801907833696483, "sagan": 0.8075590644034067, "darwin": 0.00291251275183268}], "prompt_tokens": 380, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.008937086954007608, "einstein": 0.010127046256857434, "hypatia": 0.006538512492807102, "ibn_sina": 0.0021227442136432196, "sagan": 0.9704012948891652, "socrates": 0.001873315193519514}, "confidence": 0.9020262349065657, "choice": "sagan"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 196.24237199968775, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 380, "compiled_prompt_sha256": "482a99aea3518461af305b2fde5179c25aa0874e22f2f00e33c301df075be35a", "compiled_ids_sha256": "0d8c04b699286490547a95aa4750df3f99b6a6ff4c2170fbcda9228d7de90d1c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.8125, 15.9375, 15.5, 14.375, 20.5, 14.25], "runtime_probabilities": [0.008937086910009384, 0.010127047076821327, 0.006538512650877237, 0.002122744219377637, 0.9704013466835022, 0.0018733153119683266], "decide_seconds": 0.1914967889970285, "probabilities_f64_minus_runtime_max_abs": 5.1794337041499716e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.002254958625417229, "hypatia": 0.0008295529189832139, "sagan": 0.968386829648266, "socrates": 0.002118337588820864, "darwin": 0.010757805938242028, "einstein": 0.015652515280270854}, "confidence": 0.9008769696390896, "choice": "sagan"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 194.92595299743698, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 380, "compiled_prompt_sha256": "e8f61a055e57f14754c74acd7cc3bb1a481e3083be49c4f112106963b4d19de2", "compiled_ids_sha256": "bb20ea2c7640a34981ca59f525ad9ab43740f166487587c88a26856b242b01e4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.4375, 13.4375, 20.5, 14.375, 16.0, 16.375], "runtime_probabilities": [0.0022549587301909924, 0.0008295529987663031, 0.9683868885040283, 0.0021183378994464874, 0.01075780764222145, 0.01565251685678959], "decide_seconds": 0.1904990370021551, "probabilities_f64_minus_runtime_max_abs": 5.885576237485424e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.004091381176832874, "darwin": 0.0021899585333272718, "sagan": 0.9404726886097834, "ibn_sina": 0.003610631215847928, "einstein": 0.046823378045924104, "socrates": 0.002811962418284425}, "confidence": 0.8471921118997991, "choice": "sagan"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 195.32320899816114, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 380, "compiled_prompt_sha256": "83792c4ef43956e9c0615090fdc2a9d10d767913a586f733583dd6c772ddac9f", "compiled_ids_sha256": "fb4d09450824be56ff3abbd515a120ee9004d99464b5047a0b02f36efeeaf1a2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.9375, 14.3125, 20.375, 14.8125, 17.375, 14.5625], "runtime_probabilities": [0.004091381095349789, 0.002189958468079567, 0.9404726624488831, 0.0036106312181800604, 0.046823374927043915, 0.002811962505802512], "decide_seconds": 0.1908567739992577, "probabilities_f64_minus_runtime_max_abs": 2.6160900334915027e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0029110188847721987, "einstein": 0.008423312438384236, "darwin": 0.010815747263587354, "sagan": 0.9736025415396025, "hypatia": 0.003511359035447954, "socrates": 0.0007360208382057865}, "confidence": 0.9121558624665106, "choice": "sagan"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 195.33530899934703, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 380, "compiled_prompt_sha256": "fc6239bffaf1148332bf71061007eaa3073fe882d2532f6871925929decbd17b", "compiled_ids_sha256": "42c5db44500bc5ba0789dc5d445dbc361bd1007559da29ad18c0e76fc6cb0d29", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.9375, 16.0, 16.25, 20.75, 15.125, 13.5625], "runtime_probabilities": [0.002911018906161189, 0.008423312567174435, 0.010815748013556004, 0.9736025333404541, 0.0035113589838147163, 0.0007360208546742797], "decide_seconds": 0.19095227399884607, "probabilities_f64_minus_runtime_max_abs": 8.199148360255037e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.028217289164177185, "sagan": 0.9344282837208389, "socrates": 0.010380560569090497, "ibn_sina": 0.004606345829667633, "darwin": 0.019393440319083653, "hypatia": 0.0029740803971421725}, "confidence": 0.8158150164591589, "choice": "sagan"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 195.08040100117796, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 380, "compiled_prompt_sha256": "6d488638770e470d6e8b24ecb6a22f3ac30f19d16d04d3d47158a932c3d94110", "compiled_ids_sha256": "93e3ae75a4cdf7eff2a94de5cc064b009cc7556fe7bd27a990a72760a028903a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 20.375, 15.875, 15.0625, 16.5, 14.625], "runtime_probabilities": [0.028217291459441185, 0.934428334236145, 0.010380561463534832, 0.004606346134096384, 0.019393440335989, 0.0029740806203335524], "decide_seconds": 0.1906978059996618, "probabilities_f64_minus_runtime_max_abs": 5.051530616029254e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.003980932933357408, "ibn_sina": 0.0013757736083974494, "hypatia": 0.003980932933357408, "einstein": 0.1801907833696483, "sagan": 0.8075590644034067, "darwin": 0.00291251275183268}, "confidence": 0.6922153225846583, "choice": "sagan"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 195.5100470004254, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 380, "compiled_prompt_sha256": "b0631ff6dd0c6bb01252176fe7b6fb465b6fddfad854d371cab3da4586843feb", "compiled_ids_sha256": "d352072ebd66245e2636822089274d484f0734f5444f780573b3d0a52095ed2f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.8125, 13.75, 14.8125, 18.625, 20.125, 14.5], "runtime_probabilities": [0.0039809332229197025, 0.0013757739216089249, 0.0039809332229197025, 0.1801908165216446, 0.8075591325759888, 0.0029125132132321596], "decide_seconds": 0.1910948120021203, "probabilities_f64_minus_runtime_max_abs": 6.817258202396204e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:54Z"} {"id": "c-einstein-01-dessert-02", "uid": "c020", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "prompt_sha256_by_order": ["ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf"], "request_sha256": "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "compiled_prompt_sha256_by_order": ["bf3f0bcb078d10695d6d82ba348cba3a8f6b4084459aec08e47402eef4080ca8", "d9c4f2b5904f1b737efd6cbb3b7d22702446d54d9ec55b979db7e39cbf093b36", "f3088cabb7dce69dfba27a2fb40b8b3a2bdcbefc96419894e3a8f4e9f2799ebb", "ad594906545a63fa7e9cb9afe30af90b8eb6cf2cd1d0752ba630fe842f9d2242", "f1b6ab5cfdb6fb47065c23daf89a5b247c20aeec581fc69eadb52d837b09a5d2", "a1323d1db85f8a1781b2f63f6b4cc3844dfd90bd9b75afb4402d165fe4eced3f"], "compiled_tokens_by_order": [349, 349, 349, 349, 349, 349], "choice": "sagan", "probs": {"darwin": 0.0375657552076155, "einstein": 0.04256757740179658, "hypatia": 0.04823538448364819, "ibn_sina": 0.010110685851384037, "sagan": 0.8549926476344342, "socrates": 0.006527949421121441}, "confidence": 0.6555761030137828, "seconds": 0.18662226200043128, "seconds_all_orders": 1.1197335720025876, "seconds_by_order": [0.18667653599914047, 0.1865236579978955, 0.18654036800217, 0.1866107269997883, 0.18647993800186669, 0.18690234500172664], "argmax_stable": true, "spread": {"darwin": 0.023435465156448315, "einstein": 0.3343388402363463, "hypatia": 0.037844421105530685, "ibn_sina": 0.015137308855269168, "sagan": 0.31821643601882665, "socrates": 0.017566871733567115}, "spread_2dp": {"darwin": 0.03, "einstein": 0.34, "hypatia": 0.04, "ibn_sina": 0.01, "sagan": 0.31999999999999995, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0375657552076155, "einstein": 0.04256757740179658, "hypatia": 0.04823538448364819, "ibn_sina": 0.010110685851384037, "sagan": 0.8549926476344342, "socrates": 0.006527949421121441}, {"ibn_sina": 0.018236708667069047, "hypatia": 0.010390963378117507, "sagan": 0.8254564635510782, "socrates": 0.011061122938824488, "darwin": 0.03626802584805635, "einstein": 0.09858671561685452}, {"hypatia": 0.02810143728082206, "darwin": 0.014130290051167185, "sagan": 0.7247456196340825, "ibn_sina": 0.015041615462369626, "einstein": 0.2076430965285754, "socrates": 0.010337941042983157}, {"ibn_sina": 0.020187857740661053, "einstein": 0.04273769517235377, "darwin": 0.029373159692195432, "sagan": 0.8584095543962509, "hypatia": 0.04273769517235377, "socrates": 0.0065540378261849715}, {"einstein": 0.05540065407233755, "sagan": 0.8666120377968267, "socrates": 0.01798595903785079, "ibn_sina": 0.01236155681548844, "darwin": 0.02965383323964581, "hypatia": 0.01798595903785079}, {"socrates": 0.024094821154688556, "ibn_sina": 0.005050548885391885, "hypatia": 0.03093836277316558, "einstein": 0.37690641763814287, "sagan": 0.5483956017780001, "darwin": 0.014614247770611164}], "prompt_tokens": 349, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0375657552076155, "einstein": 0.04256757740179658, "hypatia": 0.04823538448364819, "ibn_sina": 0.010110685851384037, "sagan": 0.8549926476344342, "socrates": 0.006527949421121441}, "confidence": 0.6555761030137828, "choice": "sagan"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 186.64799699763535, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 349, "compiled_prompt_sha256": "bf3f0bcb078d10695d6d82ba348cba3a8f6b4084459aec08e47402eef4080ca8", "compiled_ids_sha256": "2f7bcc4bbd9219b9afcae08151a8d821af356541e49f69a31dd887caf3fb2636", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 16.625, 16.75, 15.1875, 19.625, 14.75], "runtime_probabilities": [0.03756575658917427, 0.042567573487758636, 0.04823538661003113, 0.010110684670507908, 0.8549926280975342, 0.006527949124574661], "decide_seconds": 0.18226874099855195, "probabilities_f64_minus_runtime_max_abs": 1.9536899986682954e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.018236708667069047, "hypatia": 0.010390963378117507, "sagan": 0.8254564635510782, "socrates": 0.011061122938824488, "darwin": 0.03626802584805635, "einstein": 0.09858671561685452}, "confidence": 0.6219681665745019, "choice": "sagan"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 186.49964799988084, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 349, "compiled_prompt_sha256": "d9c4f2b5904f1b737efd6cbb3b7d22702446d54d9ec55b979db7e39cbf093b36", "compiled_ids_sha256": "92ea54ebbfdf6268cbbad32eb069488c35be4d39a71981556e1f9fd8e8069bf8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 15.375, 19.75, 15.4375, 16.625, 17.625], "runtime_probabilities": [0.018236707895994186, 0.010390963405370712, 0.825456440448761, 0.011061123572289944, 0.036268025636672974, 0.09858671575784683], "decide_seconds": 0.1824058199999854, "probabilities_f64_minus_runtime_max_abs": 2.3102317170398123e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.02810143728082206, "darwin": 0.014130290051167185, "sagan": 0.7247456196340825, "ibn_sina": 0.015041615462369626, "einstein": 0.2076430965285754, "socrates": 0.010337941042983157}, "confidence": 0.5363894330225627, "choice": "sagan"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 186.5173380028864, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 349, "compiled_prompt_sha256": "f3088cabb7dce69dfba27a2fb40b8b3a2bdcbefc96419894e3a8f4e9f2799ebb", "compiled_ids_sha256": "a08669298e97b3920d7fa38969239585ad20e8508a52b90a80fb189b725013da", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 15.5625, 19.5, 15.625, 18.25, 15.25], "runtime_probabilities": [0.02810143679380417, 0.01413029059767723, 0.7247456312179565, 0.015041614882647991, 0.20764310657978058, 0.01033794041723013], "decide_seconds": 0.18243305000214605, "probabilities_f64_minus_runtime_max_abs": 1.1583874082177203e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.020187857740661053, "einstein": 0.04273769517235377, "darwin": 0.029373159692195432, "sagan": 0.8584095543962509, "hypatia": 0.04273769517235377, "socrates": 0.0065540378261849715}, "confidence": 0.6562652779240433, "choice": "sagan"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 186.5876580013719, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 349, "compiled_prompt_sha256": "ad594906545a63fa7e9cb9afe30af90b8eb6cf2cd1d0752ba630fe842f9d2242", "compiled_ids_sha256": "d3f9b8e470ecce0a23441b6e2fa827a793bf0b5860fc115063ddf63fedc57dad", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 16.75, 16.375, 19.75, 16.75, 14.875], "runtime_probabilities": [0.020187858492136, 0.0427376925945282, 0.029373157769441605, 0.8584095239639282, 0.0427376925945282, 0.00655403733253479], "decide_seconds": 0.18237025999769685, "probabilities_f64_minus_runtime_max_abs": 3.043232266186635e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.05540065407233755, "sagan": 0.8666120377968267, "socrates": 0.01798595903785079, "ibn_sina": 0.01236155681548844, "darwin": 0.02965383323964581, "hypatia": 0.01798595903785079}, "confidence": 0.6720959626998728, "choice": "sagan"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 186.4567580014409, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 349, "compiled_prompt_sha256": "f1b6ab5cfdb6fb47065c23daf89a5b247c20aeec581fc69eadb52d837b09a5d2", "compiled_ids_sha256": "81ba9aab8a5579f40d1acd3a4a3eaed74f975055ece0a5a33d5968ccf2097c5a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 19.875, 16.0, 15.625, 16.5, 16.0], "runtime_probabilities": [0.055400654673576355, 0.8666120767593384, 0.017985958606004715, 0.0123615562915802, 0.02965383417904377, 0.017985958606004715], "decide_seconds": 0.18241005999880144, "probabilities_f64_minus_runtime_max_abs": 3.896251166146669e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.024094821154688556, "ibn_sina": 0.005050548885391885, "hypatia": 0.03093836277316558, "einstein": 0.37690641763814287, "sagan": 0.5483956017780001, "darwin": 0.014614247770611164}, "confidence": 0.45138090236386075, "choice": "sagan"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 186.8789750005817, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 349, "compiled_prompt_sha256": "a1323d1db85f8a1781b2f63f6b4cc3844dfd90bd9b75afb4402d165fe4eced3f", "compiled_ids_sha256": "51d5b5d24649adc2d94c619bd69fcb9c6532e86126594aed858286bb4bd16709", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 14.4375, 16.25, 18.75, 19.125, 15.5], "runtime_probabilities": [0.024094820022583008, 0.005050548352301121, 0.030938362702727318, 0.3769063949584961, 0.548395574092865, 0.014614246785640717], "decide_seconds": 0.18268119799904525, "probabilities_f64_minus_runtime_max_abs": 2.7685135073340916e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:56Z"} {"id": "c-einstein-01-dessert-02", "uid": "c021", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "prompt_sha256_by_order": ["dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516"], "request_sha256": "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "compiled_prompt_sha256_by_order": ["dc15cfd6dc8dccc74630a73d746dd43d7892002655d25aa889cf89abbcc1d5b6", "fec807287013fcce68866c0c36f89b0670fe2f90ae7c84a031155b0d16ad2e3e", "0434d17312dd3dd6947c54852d9e5ec5618e9f1160d8679415f29bf58ad9862d", "185f4da2619e40cd99bea0f30edf8aca1312da51d8c3ddf78bb79af6dee3dd8c", "04b7de9cd69e0223c7e901130cc94ae88621b1a4f60101fc891b6dfedd8ec3b2", "06141ca6839adc7b1fbce8f003e7b93e9e93f83ae42f445551958d390945de70"], "compiled_tokens_by_order": [287, 287, 287, 287, 287, 287], "choice": "sagan", "probs": {"darwin": 0.09899894871557191, "einstein": 0.0467637921297592, "hypatia": 0.036419677930044865, "ibn_sina": 0.07710045878293324, "sagan": 0.7315087857945192, "socrates": 0.00920833664717159}, "confidence": 0.4629491686045213, "seconds": 0.1737631904985998, "seconds_all_orders": 1.042579142991599, "seconds_by_order": [0.17409846499867854, 0.17381655699864496, 0.17368058799911523, 0.17361205799898016, 0.17371548699884443, 0.17365598799733561], "argmax_stable": true, "spread": {"darwin": 0.13630910985670935, "einstein": 0.23601186644694414, "hypatia": 0.05494728119950289, "ibn_sina": 0.0765880425876732, "sagan": 0.3192504338371861, "socrates": 0.09391271774433241}, "spread_2dp": {"darwin": 0.13, "einstein": 0.23000000000000004, "hypatia": 0.05, "ibn_sina": 0.08, "sagan": 0.32, "socrates": 0.09000000000000001}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.09899894871557191, "einstein": 0.0467637921297592, "hypatia": 0.036419677930044865, "ibn_sina": 0.07710045878293324, "sagan": 0.7315087857945192, "socrates": 0.00920833664717159}, {"ibn_sina": 0.13017270311643137, "hypatia": 0.016549602018451775, "sagan": 0.5833935808132722, "socrates": 0.03291281364076977, "darwin": 0.08946630324316346, "einstein": 0.14750499716791143}, {"hypatia": 0.07149688321795467, "darwin": 0.015953111002535807, "sagan": 0.5282944809961603, "ibn_sina": 0.09180381526580902, "einstein": 0.28277565857670334, "socrates": 0.009676050940836914}, {"ibn_sina": 0.08554153026725486, "einstein": 0.06661981075726436, "darwin": 0.058791776644057714, "sagan": 0.7162304639052606, "hypatia": 0.06661981075726436, "socrates": 0.00619660766889801}, {"einstein": 0.25004733018472425, "sagan": 0.41225835195733307, "socrates": 0.043451710615455046, "ibn_sina": 0.10423523507602299, "darwin": 0.15166137213632347, "hypatia": 0.03834600003014127}, {"socrates": 0.10010932541323042, "ibn_sina": 0.05358466052875817, "hypatia": 0.0325007395009852, "einstein": 0.14565820899946996, "sagan": 0.6527948032779421, "darwin": 0.015352262279614103}], "prompt_tokens": 287, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.09899894871557191, "einstein": 0.0467637921297592, "hypatia": 0.036419677930044865, "ibn_sina": 0.07710045878293324, "sagan": 0.7315087857945192, "socrates": 0.00920833664717159}, "confidence": 0.4629491686045213, "choice": "sagan"}}, "usage": {"input_tokens": 287, "output_tokens": 1}, "latency_ms": 174.07250500036753, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 287, "compiled_prompt_sha256": "dc15cfd6dc8dccc74630a73d746dd43d7892002655d25aa889cf89abbcc1d5b6", "compiled_ids_sha256": "770cba00074ef91bbf9b863248bc0cf872ece1572ae250a4ff112e7036365292", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 17.0, 16.75, 17.5, 19.75, 15.375], "runtime_probabilities": [0.09899894893169403, 0.046763792634010315, 0.03641967847943306, 0.07710045576095581, 0.7315087914466858, 0.009208337403833866], "decide_seconds": 0.17028940499949385, "probabilities_f64_minus_runtime_max_abs": 5.652166623626442e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.13017270311643137, "hypatia": 0.016549602018451775, "sagan": 0.5833935808132722, "socrates": 0.03291281364076977, "darwin": 0.08946630324316346, "einstein": 0.14750499716791143}, "confidence": 0.297727576133206, "choice": "sagan"}}, "usage": {"input_tokens": 287, "output_tokens": 1}, "latency_ms": 173.7932770010957, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 287, "compiled_prompt_sha256": "fec807287013fcce68866c0c36f89b0670fe2f90ae7c84a031155b0d16ad2e3e", "compiled_ids_sha256": "6a533cf70bffa70da3b7dac36b88a0245aa638de7e2abcf91ebd7100ef1c6388", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.875, 15.8125, 19.375, 16.5, 17.5, 18.0], "runtime_probabilities": [0.1301726996898651, 0.016549602150917053, 0.5833935737609863, 0.03291281685233116, 0.08946630358695984, 0.1475050002336502], "decide_seconds": 0.17032051400019554, "probabilities_f64_minus_runtime_max_abs": 7.052285866393504e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.07149688321795467, "darwin": 0.015953111002535807, "sagan": 0.5282944809961603, "ibn_sina": 0.09180381526580902, "einstein": 0.28277565857670334, "socrates": 0.009676050940836914}, "confidence": 0.3229968840929691, "choice": "sagan"}}, "usage": {"input_tokens": 287, "output_tokens": 1}, "latency_ms": 173.657857998478, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 287, "compiled_prompt_sha256": "0434d17312dd3dd6947c54852d9e5ec5618e9f1160d8679415f29bf58ad9862d", "compiled_ids_sha256": "9b367ebc79e571e8515edfe2e2a9b6c5683bc15bdd516765585784cd0351232d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 16.0, 19.5, 17.75, 18.875, 15.5], "runtime_probabilities": [0.07149688154459, 0.01595311053097248, 0.5282945036888123, 0.09180381149053574, 0.2827756702899933, 0.009676051326096058], "decide_seconds": 0.17015981599979568, "probabilities_f64_minus_runtime_max_abs": 2.2692651979738798e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.08554153026725486, "einstein": 0.06661981075726436, "darwin": 0.058791776644057714, "sagan": 0.7162304639052606, "hypatia": 0.06661981075726436, "socrates": 0.00619660766889801}, "confidence": 0.43720879664897805, "choice": "sagan"}}, "usage": {"input_tokens": 287, "output_tokens": 1}, "latency_ms": 173.58955799863907, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 287, "compiled_prompt_sha256": "185f4da2619e40cd99bea0f30edf8aca1312da51d8c3ddf78bb79af6dee3dd8c", "compiled_ids_sha256": "e63e6e5f27f1b9fc3d165a78ab8f5d3391828c2d5638850d39c7d51423bcd196", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 17.125, 17.0, 19.5, 17.125, 14.75], "runtime_probabilities": [0.08554152399301529, 0.06661980599164963, 0.05879177153110504, 0.7162303924560547, 0.06661980599164963, 0.006196607369929552], "decide_seconds": 0.17012076600076398, "probabilities_f64_minus_runtime_max_abs": 7.144920588508796e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.25004733018472425, "sagan": 0.41225835195733307, "socrates": 0.043451710615455046, "ibn_sina": 0.10423523507602299, "darwin": 0.15166137213632347, "hypatia": 0.03834600003014127}, "confidence": 0.1656516020208053, "choice": "sagan"}}, "usage": {"input_tokens": 287, "output_tokens": 1}, "latency_ms": 173.6929170001531, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 287, "compiled_prompt_sha256": "04b7de9cd69e0223c7e901130cc94ae88621b1a4f60101fc891b6dfedd8ec3b2", "compiled_ids_sha256": "923d2ecf1e4b42fbbc99c73a2362495010208919cdf1f205bf598ce4d5b2d773", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.75, 19.25, 17.0, 17.875, 18.25, 16.875], "runtime_probabilities": [0.25004732608795166, 0.4122583270072937, 0.0434517078101635, 0.10423523187637329, 0.15166138112545013, 0.03834599629044533], "decide_seconds": 0.17018476500015822, "probabilities_f64_minus_runtime_max_abs": 2.4950039367865884e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.10010932541323042, "ibn_sina": 0.05358466052875817, "hypatia": 0.0325007395009852, "einstein": 0.14565820899946996, "sagan": 0.6527948032779421, "darwin": 0.015352262279614103}, "confidence": 0.3739563778979049, "choice": "sagan"}}, "usage": {"input_tokens": 287, "output_tokens": 1}, "latency_ms": 173.63318799834815, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 287, "compiled_prompt_sha256": "06141ca6839adc7b1fbce8f003e7b93e9e93f83ae42f445551958d390945de70", "compiled_ids_sha256": "7a177e87768ffdb513b8c2dba2029ad41206a787366d8badcf71396f373a39c8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 17.0, 16.5, 18.0, 19.5, 15.75], "runtime_probabilities": [0.10010933130979538, 0.053584661334753036, 0.0325007401406765, 0.1456582248210907, 0.6527948379516602, 0.015352263115346432], "decide_seconds": 0.17015270499905455, "probabilities_f64_minus_runtime_max_abs": 3.4673718007915966e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:57Z"} {"id": "c-einstein-01-dessert-03", "uid": "c022", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "prompt_sha256_by_order": ["78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c"], "request_sha256": "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "compiled_prompt_sha256_by_order": ["7f48d90264fa637fb31af7ae8698c998de26360d2468e492bee50203c0985399", "8d80177bfd46c0d4a66807221004faa3440fa2c323586ff0418a3061b0726218", "88d56b8abe9c5079891b409265cd812ec9c063b6e44d366648d272241bca7f80", "c52b8e78a74d935c029bc1a72582b87d89615e74167dec379dce1c6fd63a3c6b", "d0d6cf9b209b4c72600a014291700cee21e8a9d81a216e2205b884caff6447b2", "ac2840d34bc92c49aaef5f31a9b4fcefea84c551148a6e40c83eb32f3a3670b0"], "compiled_tokens_by_order": [407, 407, 407, 407, 407, 407], "choice": "einstein", "probs": {"darwin": 0.014916261124306917, "einstein": 0.9228363979672984, "hypatia": 0.009047169700171782, "ibn_sina": 0.024592756994962224, "sagan": 0.024592756994962224, "socrates": 0.004014657218298278}, "confidence": 0.7857953364127328, "seconds": 0.2132079885013809, "seconds_all_orders": 1.2792479310082854, "seconds_by_order": [0.21347982800216414, 0.21301197200227762, 0.21324645999993663, 0.21333850900191464, 0.2128287330015155, 0.21334242900047684], "argmax_stable": true, "spread": {"darwin": 0.03952366283951356, "einstein": 0.05540685534981471, "hypatia": 0.02170807887942549, "ibn_sina": 0.03988283480645009, "sagan": 0.02072132373724743, "socrates": 0.008944217156068001}, "spread_2dp": {"darwin": 0.04, "einstein": 0.05999999999999994, "hypatia": 0.02, "ibn_sina": 0.04, "sagan": 0.02, "socrates": 0.01}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.014916261124306917, "einstein": 0.9228363979672984, "hypatia": 0.009047169700171782, "ibn_sina": 0.024592756994962224, "sagan": 0.024592756994962224, "socrates": 0.004014657218298278}, {"ibn_sina": 0.025244916009988523, "hypatia": 0.0032095308997375704, "sagan": 0.0038714332577147936, "socrates": 0.0030150752527169614, "darwin": 0.01735056011764368, "einstein": 0.9473084844621985}, {"hypatia": 0.01188626584399294, "darwin": 0.01048959279062084, "sagan": 0.008169303079435673, "ibn_sina": 0.02220646811200451, "einstein": 0.9442430515223232, "socrates": 0.003005318651622938}, {"ibn_sina": 0.05042826314406106, "einstein": 0.893862351937155, "darwin": 0.014447939288101425, "sagan": 0.012750261670479794, "hypatia": 0.023820624822076888, "socrates": 0.004690559138125767}, {"einstein": 0.9072280284131484, "sagan": 0.004201299272364881, "socrates": 0.006112854371532598, "ibn_sina": 0.03517704812445482, "darwin": 0.04516822387584784, "hypatia": 0.0021125459426513996}, {"socrates": 0.01194953580769094, "ibn_sina": 0.010545428337610966, "hypatia": 0.0072477598366985216, "einstein": 0.9492692072869697, "sagan": 0.015343507694695445, "darwin": 0.005644561036334286}], "prompt_tokens": 407, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.014916261124306917, "einstein": 0.9228363979672984, "hypatia": 0.009047169700171782, "ibn_sina": 0.024592756994962224, "sagan": 0.024592756994962224, "socrates": 0.004014657218298278}, "confidence": 0.7857953364127328, "choice": "einstein"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 213.44955899985507, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "7f48d90264fa637fb31af7ae8698c998de26360d2468e492bee50203c0985399", "compiled_ids_sha256": "d0e8df03971a580df0512cfa9b64c5331720d7183da1c0497ce954a5948dbfa7", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 20.625, 16.0, 17.0, 17.0, 15.1875], "runtime_probabilities": [0.014916261658072472, 0.922836422920227, 0.009047170169651508, 0.02459275722503662, 0.02459275722503662, 0.004014657344669104], "decide_seconds": 0.208581276001496, "probabilities_f64_minus_runtime_max_abs": 2.4952928612265168e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.025244916009988523, "hypatia": 0.0032095308997375704, "sagan": 0.0038714332577147936, "socrates": 0.0030150752527169614, "darwin": 0.01735056011764368, "einstein": 0.9473084844621985}, "confidence": 0.8482333549996856, "choice": "einstein"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 212.98653200210538, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "8d80177bfd46c0d4a66807221004faa3440fa2c323586ff0418a3061b0726218", "compiled_ids_sha256": "c5e49280a42ef47696d79f747609618696042a3da2bb4d64a56325287de5f62e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 14.9375, 15.125, 14.875, 16.625, 20.625], "runtime_probabilities": [0.02524491585791111, 0.0032095308415591717, 0.003871433436870575, 0.003015075344592333, 0.017350560054183006, 0.9473084807395935], "decide_seconds": 0.2083546780013421, "probabilities_f64_minus_runtime_max_abs": 3.722605002032253e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.01188626584399294, "darwin": 0.01048959279062084, "sagan": 0.008169303079435673, "ibn_sina": 0.02220646811200451, "einstein": 0.9442430515223232, "socrates": 0.003005318651622938}, "confidence": 0.8348348150824811, "choice": "einstein"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 213.22103000056813, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "88d56b8abe9c5079891b409265cd812ec9c063b6e44d366648d272241bca7f80", "compiled_ids_sha256": "28ba71f0edb6fa32de922d92e82615c9c6b6d9df91df0d03e0e636ffcbc1937e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 16.0, 15.75, 16.75, 20.5, 14.75], "runtime_probabilities": [0.011886266060173512, 0.010489593259990215, 0.008169302716851234, 0.022206468507647514, 0.9442430734634399, 0.003005318809300661], "decide_seconds": 0.20860484600052587, "probabilities_f64_minus_runtime_max_abs": 2.1941116701285068e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.05042826314406106, "einstein": 0.893862351937155, "darwin": 0.014447939288101425, "sagan": 0.012750261670479794, "hypatia": 0.023820624822076888, "socrates": 0.004690559138125767}, "confidence": 0.7310206468291344, "choice": "einstein"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 213.31349899992347, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "c52b8e78a74d935c029bc1a72582b87d89615e74167dec379dce1c6fd63a3c6b", "compiled_ids_sha256": "9b930df006e9d696ebe40b0aa154237df8c7117c0f6e84589c6dc8f79a2572c0", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 20.375, 16.25, 16.125, 16.75, 15.125], "runtime_probabilities": [0.05042826756834984, 0.8938623666763306, 0.014447939582169056, 0.012750261463224888, 0.023820625618100166, 0.0046905591152608395], "decide_seconds": 0.20865966600104002, "probabilities_f64_minus_runtime_max_abs": 1.4739175546729655e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.9072280284131484, "sagan": 0.004201299272364881, "socrates": 0.006112854371532598, "ibn_sina": 0.03517704812445482, "darwin": 0.04516822387584784, "hypatia": 0.0021125459426513996}, "confidence": 0.7694193326056579, "choice": "einstein"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 212.80261300125858, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "d0d6cf9b209b4c72600a014291700cee21e8a9d81a216e2205b884caff6447b2", "compiled_ids_sha256": "b300f63c23e343fe72534d3c1d6e785100a3ff1628049c4a29bbc4d1471ce10d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.625, 15.25, 15.625, 17.375, 17.625, 14.5625], "runtime_probabilities": [0.9072280526161194, 0.004201299510896206, 0.006112854462116957, 0.03517704829573631, 0.04516822472214699, 0.0021125460043549538], "decide_seconds": 0.2081498500010639, "probabilities_f64_minus_runtime_max_abs": 2.420297096072943e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.01194953580769094, "ibn_sina": 0.010545428337610966, "hypatia": 0.0072477598366985216, "einstein": 0.9492692072869697, "sagan": 0.015343507694695445, "darwin": 0.005644561036334286}, "confidence": 0.8440919826935451, "choice": "einstein"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 213.31668900165823, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "ac2840d34bc92c49aaef5f31a9b4fcefea84c551148a6e40c83eb32f3a3670b0", "compiled_ids_sha256": "b1c458ac84825aff6d52fdb309b3952cec0fe03f5c1727df32fcef32192f0065", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 16.25, 15.875, 20.75, 16.625, 15.625], "runtime_probabilities": [0.011949535459280014, 0.010545428842306137, 0.0072477590292692184, 0.94926917552948, 0.015343506820499897, 0.0056445603258907795], "decide_seconds": 0.2084630669996841, "probabilities_f64_minus_runtime_max_abs": 3.1757489749040246e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:58Z"} {"id": "c-einstein-01-main-01", "uid": "c023", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "prompt_sha256_by_order": ["18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c"], "request_sha256": "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "compiled_prompt_sha256_by_order": ["398a30a8df523f7677bfca21e4476fae04e1c93179ccaa5354f79989bf208257", "fb796abcfc1335e000686da4402638395dd61ebcb112ceb23cb18348e2aba549", "1b8f31e83678fceb562733937f7c4aa1690be9e6eca1423bddb061547e47f722", "3247b7295dcf1bab300f2040053fc0c5881ed636abc78e6ba4ab4ebd2b08aad7", "b3b0cf383d3ca10ad0c5f751fe4ae2e1c49ca29042b183a4ccaeaa516d4ef2cd", "293bb653e9994fc9f04a10c056b0008f872b037462957434f1ff640f7bd66ba3"], "compiled_tokens_by_order": [299, 299, 299, 299, 299, 299], "choice": "sagan", "probs": {"darwin": 0.08477745004072396, "einstein": 0.05826663249622684, "hypatia": 0.139774385157859, "ibn_sina": 0.045378099014995225, "sagan": 0.6264253342751998, "socrates": 0.045378099014995225}, "confidence": 0.31711791581554893, "seconds": 0.17512458483239848, "seconds_all_orders": 1.050747508994391, "seconds_by_order": [0.17558571299741743, 0.17475640900011058, 0.17493264799850294, 0.17498765799973626, 0.1748049189991434, 0.17568016199948033], "argmax_stable": true, "spread": {"darwin": 0.14496379303642162, "einstein": 0.29462213130595544, "hypatia": 0.19509635923215596, "ibn_sina": 0.05163904835724606, "sagan": 0.37525879485641445, "socrates": 0.09209073817713526}, "spread_2dp": {"darwin": 0.14, "einstein": 0.30000000000000004, "hypatia": 0.2, "ibn_sina": 0.05, "sagan": 0.38, "socrates": 0.09}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.08477745004072396, "einstein": 0.05826663249622684, "hypatia": 0.139774385157859, "ibn_sina": 0.045378099014995225, "sagan": 0.6264253342751998, "socrates": 0.045378099014995225}, {"ibn_sina": 0.07196041835425601, "hypatia": 0.1186426723891418, "sagan": 0.2511665394187853, "socrates": 0.1344397606854703, "darwin": 0.17262406973356126, "einstein": 0.2511665394187853}, {"hypatia": 0.1803676611603295, "darwin": 0.027660276697139625, "sagan": 0.3369711538142905, "ibn_sina": 0.06635355439306208, "einstein": 0.3369711538142905, "socrates": 0.051676200120887805}, {"ibn_sina": 0.06161746416710048, "einstein": 0.04234902250833506, "darwin": 0.06161746416710048, "sagan": 0.5159167109494873, "hypatia": 0.2761503156996416, "socrates": 0.04234902250833506}, {"einstein": 0.19443833364813592, "sagan": 0.4664333147886447, "socrates": 0.08105395646748563, "ibn_sina": 0.0433850565259024, "darwin": 0.13363538210234577, "hypatia": 0.08105395646748563}, {"socrates": 0.09107406180989537, "ibn_sina": 0.02032136999700995, "hypatia": 0.09107406180989537, "einstein": 0.17014874780326938, "sagan": 0.5938774836160923, "darwin": 0.033504274963837705}], "prompt_tokens": 299, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.08477745004072396, "einstein": 0.05826663249622684, "hypatia": 0.139774385157859, "ibn_sina": 0.045378099014995225, "sagan": 0.6264253342751998, "socrates": 0.045378099014995225}, "confidence": 0.31711791581554893, "choice": "sagan"}}, "usage": {"input_tokens": 299, "output_tokens": 1}, "latency_ms": 175.55714300033287, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 299, "compiled_prompt_sha256": "398a30a8df523f7677bfca21e4476fae04e1c93179ccaa5354f79989bf208257", "compiled_ids_sha256": "4e8ff629b9c571ecae098a26e5b2d5c09709bf9dbea7851caa35fd43fed66dce", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.25, 16.875, 17.75, 16.625, 19.25, 16.625], "runtime_probabilities": [0.08477744460105896, 0.05826662853360176, 0.1397743821144104, 0.045378100126981735, 0.6264253258705139, 0.045378100126981735], "decide_seconds": 0.17175139299797593, "probabilities_f64_minus_runtime_max_abs": 8.404685836183035e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.07196041835425601, "hypatia": 0.1186426723891418, "sagan": 0.2511665394187853, "socrates": 0.1344397606854703, "darwin": 0.17262406973356126, "einstein": 0.2511665394187853}, "confidence": 0.04600450597272254, "choice": "sagan"}}, "usage": {"input_tokens": 299, "output_tokens": 1}, "latency_ms": 174.73283900108072, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 299, "compiled_prompt_sha256": "fb796abcfc1335e000686da4402638395dd61ebcb112ceb23cb18348e2aba549", "compiled_ids_sha256": "89d17dbea35c0b40e89d1602d364d14a2483e687edf0a8b96361ac1944d79cb8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 17.5, 18.25, 17.625, 17.875, 18.25], "runtime_probabilities": [0.07196041941642761, 0.11864268034696579, 0.25116655230522156, 0.13443976640701294, 0.17262408137321472, 0.25116655230522156], "decide_seconds": 0.17109200800041435, "probabilities_f64_minus_runtime_max_abs": 1.2886436251413613e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.1803676611603295, "darwin": 0.027660276697139625, "sagan": 0.3369711538142905, "ibn_sina": 0.06635355439306208, "einstein": 0.3369711538142905, "socrates": 0.051676200120887805}, "confidence": 0.1771463437543399, "choice": "sagan"}}, "usage": {"input_tokens": 299, "output_tokens": 1}, "latency_ms": 174.90881799676572, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 299, "compiled_prompt_sha256": "1b8f31e83678fceb562733937f7c4aa1690be9e6eca1423bddb061547e47f722", "compiled_ids_sha256": "7f77fd5c86cfca37884acf2197ce451c9662901c9e6acd77875110527e99ebec", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 16.25, 18.75, 17.125, 18.75, 16.875], "runtime_probabilities": [0.18036767840385437, 0.027660276740789413, 0.3369711637496948, 0.06635355204343796, 0.3369711637496948, 0.051676202565431595], "decide_seconds": 0.17121259700070368, "probabilities_f64_minus_runtime_max_abs": 1.7243524857990522e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.06161746416710048, "einstein": 0.04234902250833506, "darwin": 0.06161746416710048, "sagan": 0.5159167109494873, "hypatia": 0.2761503156996416, "socrates": 0.04234902250833506}, "confidence": 0.2699783400361814, "choice": "sagan"}}, "usage": {"input_tokens": 299, "output_tokens": 1}, "latency_ms": 174.96418800146785, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 299, "compiled_prompt_sha256": "3247b7295dcf1bab300f2040053fc0c5881ed636abc78e6ba4ab4ebd2b08aad7", "compiled_ids_sha256": "267064f650ecc4e24294fcecd4900507c9af3237fcd9d02a31158c23fb67dd9f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 16.5, 16.875, 19.0, 18.375, 16.5], "runtime_probabilities": [0.06161746010184288, 0.04234901815652847, 0.06161746010184288, 0.515916645526886, 0.27615031599998474, 0.04234901815652847], "decide_seconds": 0.17131196599802934, "probabilities_f64_minus_runtime_max_abs": 6.542260133635125e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.19443833364813592, "sagan": 0.4664333147886447, "socrates": 0.08105395646748563, "ibn_sina": 0.0433850565259024, "darwin": 0.13363538210234577, "hypatia": 0.08105395646748563}, "confidence": 0.17034268894206173, "choice": "sagan"}}, "usage": {"input_tokens": 299, "output_tokens": 1}, "latency_ms": 174.7801090023131, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 299, "compiled_prompt_sha256": "b3b0cf383d3ca10ad0c5f751fe4ae2e1c49ca29042b183a4ccaeaa516d4ef2cd", "compiled_ids_sha256": "00b244220fca4faeb874f282a9c1621ba27fca6bedca2049e1f9f71a780b506a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 19.0, 17.25, 16.625, 17.75, 17.25], "runtime_probabilities": [0.19443835318088531, 0.4664333462715149, 0.0810539573431015, 0.043385058641433716, 0.13363540172576904, 0.0810539573431015], "decide_seconds": 0.17110188800143078, "probabilities_f64_minus_runtime_max_abs": 3.1482870199717183e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.09107406180989537, "ibn_sina": 0.02032136999700995, "hypatia": 0.09107406180989537, "einstein": 0.17014874780326938, "sagan": 0.5938774836160923, "darwin": 0.033504274963837705}, "confidence": 0.307828232788234, "choice": "sagan"}}, "usage": {"input_tokens": 299, "output_tokens": 1}, "latency_ms": 175.64934199981508, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 299, "compiled_prompt_sha256": "293bb653e9994fc9f04a10c056b0008f872b037462957434f1ff640f7bd66ba3", "compiled_ids_sha256": "f9154e3a4b7bc6bab19845d1da15fc6951a0f6caf7ce89a37aab4e51345472e6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 15.875, 17.375, 18.0, 19.25, 16.375], "runtime_probabilities": [0.09107405692338943, 0.020321369171142578, 0.09107405692338943, 0.17014874517917633, 0.5938774347305298, 0.03350427374243736], "decide_seconds": 0.17181571199762402, "probabilities_f64_minus_runtime_max_abs": 4.888556248694442e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:36:59Z"} {"id": "c-einstein-01-main-01", "uid": "c024", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "prompt_sha256_by_order": ["0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2"], "request_sha256": "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "compiled_prompt_sha256_by_order": ["31ad7358200e1676880e3e7daab4375eb6f0352b2bc77278e208e561783c83f2", "e0c2f8e76dfd6c1703219e9e816b92c0098057c8e124d4b6a0f6d23a741ad3fc", "d1e86d115749e78a173b8a21a6b94e16f0e4d7bb92c0f51751a4ac82a7a313fa", "47a9372a4aa4f3f9bf55c5958e2d6a15407f66a6330504bd4ed931d10e3aa32a", "9703b21f7fe24b0adf4d457b162e354446267f4af9f4a26da7f51d3dbbde02a8", "b4f413074fb81b8da1b697c2898480d0882dfbb137eee1f36bcb897aabcb1905"], "compiled_tokens_by_order": [346, 346, 346, 346, 346, 346], "choice": "einstein", "probs": {"darwin": 0.004462998399070655, "einstein": 0.9637393044177925, "hypatia": 0.0021081711686576727, "ibn_sina": 0.015577395031359733, "sagan": 0.012131687448635572, "socrates": 0.001980443534483769}, "confidence": 0.8864677420128372, "seconds": 0.18638558399955704, "seconds_all_orders": 1.1183135039973422, "seconds_by_order": [0.1862912199976563, 0.18629710000095656, 0.1861726200004341, 0.18640549900010228, 0.18664185699890368, 0.18650520799928927], "argmax_stable": true, "spread": {"darwin": 0.012248943266093049, "einstein": 0.03907918655435727, "hypatia": 0.006384957992434747, "ibn_sina": 0.01757774864452582, "sagan": 0.015034091497126596, "socrates": 0.007943396763450222}, "spread_2dp": {"darwin": 0.01, "einstein": 0.040000000000000036, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.02, "socrates": 0.01}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.004462998399070655, "einstein": 0.9637393044177925, "hypatia": 0.0021081711686576727, "ibn_sina": 0.015577395031359733, "sagan": 0.012131687448635572, "socrates": 0.001980443534483769}, {"ibn_sina": 0.02498116499649232, "hypatia": 0.004341075553426695, "sagan": 0.019455350861304187, "socrates": 0.004621050872366497, "darwin": 0.00919005701872119, "einstein": 0.9374113006976892}, {"hypatia": 0.007371474302421438, "darwin": 0.004200139450465097, "sagan": 0.006111166840625927, "ibn_sina": 0.013771727065814364, "einstein": 0.9654726046749577, "socrates": 0.0030728876657156244}, {"ibn_sina": 0.008448298064451461, "einstein": 0.9764904872520465, "darwin": 0.0029196536860883774, "sagan": 0.005454631196136699, "hypatia": 0.005124151798480709, "socrates": 0.0015627780027964006}, {"einstein": 0.9547261825526161, "sagan": 0.004421259364177591, "socrates": 0.006432894416678723, "ibn_sina": 0.019814709769564533, "darwin": 0.01361843758697644, "hypatia": 0.000986516309986691}, {"socrates": 0.009506174766246623, "ibn_sina": 0.007403416351966499, "hypatia": 0.0025585522287167816, "einstein": 0.96965618756594, "sagan": 0.009506174766246623, "darwin": 0.0013694943208833908}], "prompt_tokens": 346, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.004462998399070655, "einstein": 0.9637393044177925, "hypatia": 0.0021081711686576727, "ibn_sina": 0.015577395031359733, "sagan": 0.012131687448635572, "socrates": 0.001980443534483769}, "confidence": 0.8864677420128372, "choice": "einstein"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 186.25620999955572, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 346, "compiled_prompt_sha256": "31ad7358200e1676880e3e7daab4375eb6f0352b2bc77278e208e561783c83f2", "compiled_ids_sha256": "c9faf3bf2ed711f2c81c34c713922081a5f5829813476075c406186f3052df81", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5, 20.875, 14.75, 16.75, 16.5, 14.6875], "runtime_probabilities": [0.004462998826056719, 0.9637393355369568, 0.002108171349391341, 0.01557739544659853, 0.012131688185036182, 0.001980443485081196], "decide_seconds": 0.18201715299801435, "probabilities_f64_minus_runtime_max_abs": 3.111916424547445e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.02498116499649232, "hypatia": 0.004341075553426695, "sagan": 0.019455350861304187, "socrates": 0.004621050872366497, "darwin": 0.00919005701872119, "einstein": 0.9374113006976892}, "confidence": 0.8208654794431881, "choice": "einstein"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 186.26990000120713, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 346, "compiled_prompt_sha256": "e0c2f8e76dfd6c1703219e9e816b92c0098057c8e124d4b6a0f6d23a741ad3fc", "compiled_ids_sha256": "8f7149783acf5f26b2416319ec825186cc44b96cf2da6456b8aacab1480206b9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 15.375, 16.875, 15.4375, 16.125, 20.75], "runtime_probabilities": [0.024981165304780006, 0.004341076128184795, 0.019455350935459137, 0.004621050786226988, 0.009190057404339314, 0.9374113082885742], "decide_seconds": 0.18214660200101207, "probabilities_f64_minus_runtime_max_abs": 7.590885031660832e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.007371474302421438, "darwin": 0.004200139450465097, "sagan": 0.006111166840625927, "ibn_sina": 0.013771727065814364, "einstein": 0.9654726046749577, "socrates": 0.0030728876657156244}, "confidence": 0.8877927105007091, "choice": "einstein"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 186.14641100066365, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 346, "compiled_prompt_sha256": "d1e86d115749e78a173b8a21a6b94e16f0e4d7bb92c0f51751a4ac82a7a313fa", "compiled_ids_sha256": "ba200ceb0620aa31241c2cd17a3d157bc37d9433ee04195900e2e5b7a11ef412", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.75, 15.1875, 15.5625, 16.375, 20.625, 14.875], "runtime_probabilities": [0.0073714726604521275, 0.004200139082968235, 0.006111165974289179, 0.01377172488719225, 0.9654724597930908, 0.0030728874262422323], "decide_seconds": 0.18202927300080773, "probabilities_f64_minus_runtime_max_abs": 1.448818668370322e-07}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.008448298064451461, "einstein": 0.9764904872520465, "darwin": 0.0029196536860883774, "sagan": 0.005454631196136699, "hypatia": 0.005124151798480709, "socrates": 0.0015627780027964006}, "confidence": 0.9184331069980567, "choice": "einstein"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 186.37881899849162, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 346, "compiled_prompt_sha256": "47a9372a4aa4f3f9bf55c5958e2d6a15407f66a6330504bd4ed931d10e3aa32a", "compiled_ids_sha256": "f7003fc8d27823fc8645d3ec99b5a66135b4cf57b491b6aee4da2a8691d0a385", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 21.0, 15.1875, 15.8125, 15.75, 14.5625], "runtime_probabilities": [0.008448298089206219, 0.9764904975891113, 0.00291965389624238, 0.005454631522297859, 0.005124151706695557, 0.001562778023071587], "decide_seconds": 0.18227600099999108, "probabilities_f64_minus_runtime_max_abs": 1.0337064870746815e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.9547261825526161, "sagan": 0.004421259364177591, "socrates": 0.006432894416678723, "ibn_sina": 0.019814709769564533, "darwin": 0.01361843758697644, "hypatia": 0.000986516309986691}, "confidence": 0.8639872680805171, "choice": "einstein"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 186.61496699860436, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 346, "compiled_prompt_sha256": "9703b21f7fe24b0adf4d457b162e354446267f4af9f4a26da7f51d3dbbde02a8", "compiled_ids_sha256": "7c92edd32eeb651990329eadde74066cb997508e2801de62eab97e40a22c8ff1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.875, 15.5, 15.875, 17.0, 16.625, 14.0], "runtime_probabilities": [0.9547261595726013, 0.004421259742230177, 0.006432894617319107, 0.019814709201455116, 0.013618436641991138, 0.0009865162428468466], "decide_seconds": 0.18240924999918207, "probabilities_f64_minus_runtime_max_abs": 2.298001477996081e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.009506174766246623, "ibn_sina": 0.007403416351966499, "hypatia": 0.0025585522287167816, "einstein": 0.96965618756594, "sagan": 0.009506174766246623, "darwin": 0.0013694943208833908}, "confidence": 0.9000891577018832, "choice": "einstein"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 186.47823799983598, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 346, "compiled_prompt_sha256": "b4f413074fb81b8da1b697c2898480d0882dfbb137eee1f36bcb897aabcb1905", "compiled_ids_sha256": "3898042e21451135a44b30cd6af9147537d0ac85cb9072a829ace1430945ff80", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 16.125, 15.0625, 21.0, 16.375, 14.4375], "runtime_probabilities": [0.009506176225841045, 0.007403416559100151, 0.0025585524272173643, 0.9696562886238098, 0.009506176225841045, 0.00136949447914958], "decide_seconds": 0.18220077200021478, "probabilities_f64_minus_runtime_max_abs": 1.0105786985015186e-07}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:00Z"} {"id": "c-einstein-01-main-02", "uid": "c025", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "prompt_sha256_by_order": ["af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11"], "request_sha256": "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "compiled_prompt_sha256_by_order": ["8bc0000acb4ce0e5f74190578a430fa97ca4b3961d56af5f977ae0c41db74aee", "0494bc9fefd8d4c25ac86fdcaef4fbc76617fddf7ec9a239458b2053f77e14e5", "76ed487276a05ea6bf05d42b96ee8cd806fd097a7505da0d32b0d4a3b181c789", "00f01aac12f5e659cbc7bc3aba83f2121260446d262a1b7f39663aff3aff6f3b", "40382364013268e8833268571f4b4cceee955f64fe25674b6d23848afe4523f1", "7d295551f16192316aaa0e865f85be295172fe2d1500426daebd57fa6cb99971"], "compiled_tokens_by_order": [384, 384, 384, 384, 384, 384], "choice": "sagan", "probs": {"darwin": 0.014514889752784612, "einstein": 0.2572825769443648, "hypatia": 0.012809345248189302, "ibn_sina": 0.006050706257756679, "sagan": 0.6993665536869829, "socrates": 0.00997592810992181}, "confidence": 0.5571481577290924, "seconds": 0.19506644333402315, "seconds_all_orders": 1.170398660004139, "seconds_by_order": [0.19533263000266743, 0.1950957510016451, 0.19488401299895486, 0.19471455399980186, 0.19522767099988414, 0.19514404100118554], "argmax_stable": false, "spread": {"darwin": 0.02959685937962224, "einstein": 0.6235954872653837, "hypatia": 0.012740657069292427, "ibn_sina": 0.011433001519930678, "sagan": 0.6026951176613891, "socrates": 0.0387473262204944}, "spread_2dp": {"darwin": 0.03, "einstein": 0.6299999999999999, "hypatia": 0.02, "ibn_sina": 0.01, "sagan": 0.6000000000000001, "socrates": 0.04}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.014514889752784612, "einstein": 0.2572825769443648, "hypatia": 0.012809345248189302, "ibn_sina": 0.006050706257756679, "sagan": 0.6993665536869829, "socrates": 0.00997592810992181}, {"ibn_sina": 0.010275949222594505, "hypatia": 0.003551273025279211, "sagan": 0.7203996575036143, "socrates": 0.014951417895927585, "darwin": 0.016942176059926005, "einstein": 0.23387952629265846}, {"hypatia": 0.009503173142328753, "darwin": 0.002120444545166072, "sagan": 0.48741979694593135, "ibn_sina": 0.00613570973573368, "einstein": 0.48741979694593135, "socrates": 0.0074010786849087766}, {"ibn_sina": 0.011197268884811078, "einstein": 0.06443583348848952, "darwin": 0.009282861949603764, "sagan": 0.8895092436329203, "hypatia": 0.01629193009457164, "socrates": 0.009282861949603764}, {"einstein": 0.34099315272219044, "sagan": 0.5622026640561727, "socrates": 0.04614840490540317, "ibn_sina": 0.01407446829964889, "darwin": 0.03171730392478831, "hypatia": 0.00486400609179643}, {"socrates": 0.014279634498613768, "ibn_sina": 0.0026414667797182117, "hypatia": 0.005591985216545282, "einstein": 0.6880313207538733, "sagan": 0.2868141259715312, "darwin": 0.0026414667797182117}], "prompt_tokens": 384, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.014514889752784612, "einstein": 0.2572825769443648, "hypatia": 0.012809345248189302, "ibn_sina": 0.006050706257756679, "sagan": 0.6993665536869829, "socrates": 0.00997592810992181}, "confidence": 0.5571481577290924, "choice": "sagan"}}, "usage": {"input_tokens": 384, "output_tokens": 1}, "latency_ms": 195.30282999767223, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 384, "compiled_prompt_sha256": "8bc0000acb4ce0e5f74190578a430fa97ca4b3961d56af5f977ae0c41db74aee", "compiled_ids_sha256": "96c01d22640684d10ebbdf561b48b5676cfc3b4bf58f2c5ccd9550d3ab256a3e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.875, 18.75, 15.75, 15.0, 19.75, 15.5], "runtime_probabilities": [0.014514889568090439, 0.2572826147079468, 0.012809345498681068, 0.006050706375390291, 0.699366569519043, 0.009975927881896496], "decide_seconds": 0.1906895149986667, "probabilities_f64_minus_runtime_max_abs": 3.776358198370744e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.010275949222594505, "hypatia": 0.003551273025279211, "sagan": 0.7203996575036143, "socrates": 0.014951417895927585, "darwin": 0.016942176059926005, "einstein": 0.23387952629265846}, "confidence": 0.5674238377357768, "choice": "sagan"}}, "usage": {"input_tokens": 384, "output_tokens": 1}, "latency_ms": 195.07019100274192, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 384, "compiled_prompt_sha256": "0494bc9fefd8d4c25ac86fdcaef4fbc76617fddf7ec9a239458b2053f77e14e5", "compiled_ids_sha256": "4a8956bac8ac84f21e010cb16becbc7f4be64a5c282e86a2d016ebd22a3d9d59", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.625, 14.5625, 19.875, 16.0, 16.125, 18.75], "runtime_probabilities": [0.010275949724018574, 0.003551273141056299, 0.7203997373580933, 0.014951419085264206, 0.016942176967859268, 0.23387953639030457], "decide_seconds": 0.1904300569985935, "probabilities_f64_minus_runtime_max_abs": 7.985447891822872e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.009503173142328753, "darwin": 0.002120444545166072, "sagan": 0.48741979694593135, "ibn_sina": 0.00613570973573368, "einstein": 0.48741979694593135, "socrates": 0.0074010786849087766}, "confidence": 0.5393277145235251, "choice": "sagan"}}, "usage": {"input_tokens": 384, "output_tokens": 1}, "latency_ms": 194.8588129998825, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 384, "compiled_prompt_sha256": "76ed487276a05ea6bf05d42b96ee8cd806fd097a7505da0d32b0d4a3b181c789", "compiled_ids_sha256": "b3bf1be97c87930b57ae8957c6dc8281df31600a11862e149815ef7e9466c54f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.4375, 13.9375, 19.375, 15.0, 19.375, 15.1875], "runtime_probabilities": [0.009503173641860485, 0.0021204445511102676, 0.48741981387138367, 0.006135710049420595, 0.48741981387138367, 0.007401078473776579], "decide_seconds": 0.19036128699735855, "probabilities_f64_minus_runtime_max_abs": 1.692545231746223e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.011197268884811078, "einstein": 0.06443583348848952, "darwin": 0.009282861949603764, "sagan": 0.8895092436329203, "hypatia": 0.01629193009457164, "socrates": 0.009282861949603764}, "confidence": 0.7292653739157215, "choice": "sagan"}}, "usage": {"input_tokens": 384, "output_tokens": 1}, "latency_ms": 194.69007400039118, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 384, "compiled_prompt_sha256": "00f01aac12f5e659cbc7bc3aba83f2121260446d262a1b7f39663aff3aff6f3b", "compiled_ids_sha256": "43d9a73f26bee34e62258a74f4dfa19b595076038387a93a74238d4108528dae", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.75, 17.5, 15.5625, 20.125, 16.125, 15.5625], "runtime_probabilities": [0.011197268031537533, 0.06443583220243454, 0.009282861836254597, 0.8895092010498047, 0.01629192940890789, 0.009282861836254597], "decide_seconds": 0.1902642079985526, "probabilities_f64_minus_runtime_max_abs": 4.2583115589245324e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.34099315272219044, "sagan": 0.5622026640561727, "socrates": 0.04614840490540317, "ibn_sina": 0.01407446829964889, "darwin": 0.03171730392478831, "hypatia": 0.00486400609179643}, "confidence": 0.4262894279353058, "choice": "sagan"}}, "usage": {"input_tokens": 384, "output_tokens": 1}, "latency_ms": 195.20163099878118, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 384, "compiled_prompt_sha256": "40382364013268e8833268571f4b4cceee955f64fe25674b6d23848afe4523f1", "compiled_ids_sha256": "3949bd5cec5d7bf94d9f984527eb8d1d4729b1390d3aab6fe96ef9bd24b0e01d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.125, 19.625, 17.125, 15.9375, 16.75, 14.875], "runtime_probabilities": [0.34099313616752625, 0.5622026324272156, 0.046148400753736496, 0.014074468053877354, 0.03171730414032936, 0.004864005837589502], "decide_seconds": 0.19058182599837892, "probabilities_f64_minus_runtime_max_abs": 3.1628957120055645e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.014279634498613768, "ibn_sina": 0.0026414667797182117, "hypatia": 0.005591985216545282, "einstein": 0.6880313207538733, "sagan": 0.2868141259715312, "darwin": 0.0026414667797182117}, "confidence": 0.5889433704307667, "choice": "einstein"}}, "usage": {"input_tokens": 384, "output_tokens": 1}, "latency_ms": 195.1188610000827, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 384, "compiled_prompt_sha256": "7d295551f16192316aaa0e865f85be295172fe2d1500426daebd57fa6cb99971", "compiled_ids_sha256": "2cce23a5300473a30b8b4c1d5991f9b53b5b8a138d40cadb5e63259074bbd485", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 14.3125, 15.0625, 19.875, 19.0, 14.3125], "runtime_probabilities": [0.014279632829129696, 0.0026414666790515184, 0.0055919853039085865, 0.6880312561988831, 0.2868140935897827, 0.0026414666790515184], "decide_seconds": 0.1904760770012217, "probabilities_f64_minus_runtime_max_abs": 6.455499024227151e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:01Z"} {"id": "c-einstein-01-main-03", "uid": "c026", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "prompt_sha256_by_order": ["fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd"], "request_sha256": "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "compiled_prompt_sha256_by_order": ["babae6f4edd94f8ec21e9cf611f98b3fccee444a918dac8ccf6454902374a632", "1184b761fd135b39e09455ad71e100141129446e3b27aa3aaec8a1e9e2f23378", "48e09c0736f5871ec68014b23b8b65dcbfdf6135a77b7fe34ef3001271e60c91", "129ed980177a61906be57f39da63f7afb936fdd51f73e537dea1d40e65b51681", "f6768315e2c58c3e890e1ba6bc6ab8f3937fcbf10b7475df64e2b254f1fa389b", "d76a2f91dac671039feb16b900913661869267184bfad121f33add519ad4a47f"], "compiled_tokens_by_order": [339, 339, 339, 339, 339, 339], "choice": "sagan", "probs": {"darwin": 0.00953196832362797, "einstein": 0.004502583017868693, "hypatia": 0.006154301302107744, "ibn_sina": 0.0032941601065024726, "sagan": 0.9722872019465052, "socrates": 0.0042297853033879725}, "confidence": 0.9055219595195507, "seconds": 0.18623999816615347, "seconds_all_orders": 1.1174399889969209, "seconds_by_order": [0.18626094999854104, 0.1860783709998941, 0.18673234599918942, 0.18649378799818805, 0.18644451799991657, 0.1854300160011917], "argmax_stable": true, "spread": {"darwin": 0.007399863708765135, "einstein": 0.009259428979779973, "hypatia": 0.005878919070934764, "ibn_sina": 0.0014650806135225554, "sagan": 0.015503595663795222, "socrates": 0.00582689303125088}, "spread_2dp": {"darwin": 0.01, "einstein": 0.01, "hypatia": 0.01, "ibn_sina": 0.0, "sagan": 0.020000000000000018, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.00953196832362797, "einstein": 0.004502583017868693, "hypatia": 0.006154301302107744, "ibn_sina": 0.0032941601065024726, "sagan": 0.9722872019465052, "socrates": 0.0042297853033879725}, {"ibn_sina": 0.0022784298136838466, "hypatia": 0.00156594038342262, "sagan": 0.9784664778233546, "socrates": 0.0022784298136838466, "darwin": 0.005818174353159349, "einstein": 0.009592547812695777}, {"hypatia": 0.0033022682273572183, "darwin": 0.0021321046148628352, "sagan": 0.9746803528207882, "ibn_sina": 0.0031021939096932717, "einstein": 0.012269414922213838, "socrates": 0.004513665505084582}, {"ibn_sina": 0.003743510427206402, "einstein": 0.006570065408648033, "darwin": 0.0042419530496284686, "sagan": 0.9750841628080155, "hypatia": 0.007444859454357384, "socrates": 0.0029154488521443154}, {"einstein": 0.005475883708900794, "sagan": 0.9802951214790205, "socrates": 0.0020144650387502287, "ibn_sina": 0.0024299086661415344, "darwin": 0.007031173861054957, "hypatia": 0.002753447246131955}, {"socrates": 0.007841358070001109, "ibn_sina": 0.0034795815509455603, "hypatia": 0.0050627612830896765, "einstein": 0.013762011997648666, "sagan": 0.9647915258152253, "darwin": 0.0050627612830896765}], "prompt_tokens": 339, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.00953196832362797, "einstein": 0.004502583017868693, "hypatia": 0.006154301302107744, "ibn_sina": 0.0032941601065024726, "sagan": 0.9722872019465052, "socrates": 0.0042297853033879725}, "confidence": 0.9055219595195507, "choice": "sagan"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 186.22907999815652, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "babae6f4edd94f8ec21e9cf611f98b3fccee444a918dac8ccf6454902374a632", "compiled_ids_sha256": "6866bdd0a49c4b3aa6a808c4deba4a1c9165cab5449d0a129281ece86c4bd505", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.75, 15.0, 15.3125, 14.6875, 20.375, 14.9375], "runtime_probabilities": [0.009531970135867596, 0.004502583760768175, 0.006154301576316357, 0.0032941605895757675, 0.9722872972488403, 0.0042297858744859695], "decide_seconds": 0.1819491240021307, "probabilities_f64_minus_runtime_max_abs": 9.530233513466158e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0022784298136838466, "hypatia": 0.00156594038342262, "sagan": 0.9784664778233546, "socrates": 0.0022784298136838466, "darwin": 0.005818174353159349, "einstein": 0.009592547812695777}, "confidence": 0.9254034184438609, "choice": "sagan"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 186.04969099760638, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "1184b761fd135b39e09455ad71e100141129446e3b27aa3aaec8a1e9e2f23378", "compiled_ids_sha256": "463355481f2dffe0a9c016bddd3af0e00894e36f2d49f66a397429da5cacae06", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.5625, 14.1875, 20.625, 14.5625, 15.5, 16.0], "runtime_probabilities": [0.002278429688885808, 0.001565940328873694, 0.9784664511680603, 0.002278429688885808, 0.0058181737549602985, 0.00959254801273346], "decide_seconds": 0.18191112400018028, "probabilities_f64_minus_runtime_max_abs": 2.6655294305122368e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0033022682273572183, "darwin": 0.0021321046148628352, "sagan": 0.9746803528207882, "ibn_sina": 0.0031021939096932717, "einstein": 0.012269414922213838, "socrates": 0.004513665505084582}, "confidence": 0.9144618704162443, "choice": "sagan"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 186.70284600011655, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "48e09c0736f5871ec68014b23b8b65dcbfdf6135a77b7fe34ef3001271e60c91", "compiled_ids_sha256": "4e03e8345f39d057cb29a3a6063270e2508890706a054f5d717d30d2f3680912", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.9375, 14.5, 20.625, 14.875, 16.25, 15.25], "runtime_probabilities": [0.0033022682182490826, 0.0021321047097444534, 0.9746803641319275, 0.0031021940521895885, 0.012269414961338043, 0.004513666033744812], "decide_seconds": 0.18215499199868646, "probabilities_f64_minus_runtime_max_abs": 1.1311139247460744e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.003743510427206402, "einstein": 0.006570065408648033, "darwin": 0.0042419530496284686, "sagan": 0.9750841628080155, "hypatia": 0.007444859454357384, "socrates": 0.0029154488521443154}, "confidence": 0.9133753981259184, "choice": "sagan"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 186.46523800271098, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "129ed980177a61906be57f39da63f7afb936fdd51f73e537dea1d40e65b51681", "compiled_ids_sha256": "05a1442043d0408e3840b5fd47a230f9ec1e5f7740b118681f2f7ee04f65e15d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.1875, 15.75, 15.3125, 20.75, 15.875, 14.9375], "runtime_probabilities": [0.003743510926142335, 0.006570066325366497, 0.004241953603923321, 0.9750842452049255, 0.0074448599480092525, 0.002915449207648635], "decide_seconds": 0.18205745299928822, "probabilities_f64_minus_runtime_max_abs": 8.239691007272398e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.005475883708900794, "sagan": 0.9802951214790205, "socrates": 0.0020144650387502287, "ibn_sina": 0.0024299086661415344, "darwin": 0.007031173861054957, "hypatia": 0.002753447246131955}, "confidence": 0.9295415287999295, "choice": "sagan"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 186.41648899938446, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "f6768315e2c58c3e890e1ba6bc6ab8f3937fcbf10b7475df64e2b254f1fa389b", "compiled_ids_sha256": "146b49a177ff9a096200876aba7359d1ef49602baa41a718aaf8dcff1296c6eb", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.6875, 20.875, 14.6875, 14.875, 15.9375, 15.0], "runtime_probabilities": [0.005475883837789297, 0.9802951216697693, 0.002014464931562543, 0.0024299088399857283, 0.007031173445284367, 0.0027534475084394217], "decide_seconds": 0.18202145299801487, "probabilities_f64_minus_runtime_max_abs": 4.157705906990272e-10}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.007841358070001109, "ibn_sina": 0.0034795815509455603, "hypatia": 0.0050627612830896765, "einstein": 0.013762011997648666, "sagan": 0.9647915258152253, "darwin": 0.0050627612830896765}, "confidence": 0.885698964441118, "choice": "sagan"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 185.402865998185, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "d76a2f91dac671039feb16b900913661869267184bfad121f33add519ad4a47f", "compiled_ids_sha256": "776c074e939bce142305e4890110d00b065f7e733c091c93a6373e0fe023172f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.4375, 14.625, 15.0, 16.0, 20.25, 15.0], "runtime_probabilities": [0.00784135889261961, 0.003479581791907549, 0.005062761716544628, 0.013762012124061584, 0.9647915959358215, 0.005062761716544628], "decide_seconds": 0.18131335900034173, "probabilities_f64_minus_runtime_max_abs": 7.012059621835931e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:02Z"} {"id": "c-einstein-01-main-03", "uid": "c027", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "prompt_sha256_by_order": ["7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af"], "request_sha256": "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "compiled_prompt_sha256_by_order": ["5b0777e33fab5b56af64797557c932888e6e56a8386aa298f23774efd807c843", "0d097107b6a8fe4f153aa29a3af80c9c138fd5d2816bebbbec8d60ed14143d0a", "08504ba6b8e244df702dd952429002075c657f9019924174db85f519dc7d4195", "975fb5a80fe63ff6e9b0cea5774e58f849b0b863b1fb38437460dafb81c77b98", "0479b973affcdcbaddd3c7ab9b3104ed54e2f719ac405e19ef024dec2504ae60", "c3c12676b722efd1aae12c93922ffc374ad0ba0d69b50bbaf70f9d7e70b298bf"], "compiled_tokens_by_order": [345, 345, 345, 345, 345, 345], "choice": "einstein", "probs": {"darwin": 0.0031138858688066727, "einstein": 0.9783538571746611, "hypatia": 0.002140140373009538, "ibn_sina": 0.005133929866434111, "sagan": 0.009591443718706075, "socrates": 0.0016667429983825567}, "confidence": 0.9247465361875539, "seconds": 0.18603987483341675, "seconds_all_orders": 1.1162392490005004, "seconds_by_order": [0.18612490099985735, 0.1858266929993988, 0.18580874299732386, 0.1863980790003552, 0.1861973510021926, 0.1858834820013726], "argmax_stable": true, "spread": {"darwin": 0.007294605898142287, "einstein": 0.03744431852015506, "hypatia": 0.008385086819650256, "ibn_sina": 0.01785460903197638, "sagan": 0.016987881916643743, "socrates": 0.005295116269640509}, "spread_2dp": {"darwin": 0.01, "einstein": 0.040000000000000036, "hypatia": 0.01, "ibn_sina": 0.02, "sagan": 0.02, "socrates": 0.01}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0031138858688066727, "einstein": 0.9783538571746611, "hypatia": 0.002140140373009538, "ibn_sina": 0.005133929866434111, "sagan": 0.009591443718706075, "socrates": 0.0016667429983825567}, {"ibn_sina": 0.022128071403571446, "hypatia": 0.005955689622689923, "sagan": 0.019527954473822565, "socrates": 0.002813268576265473, "darwin": 0.008665477269144553, "einstein": 0.940909538654506}, {"hypatia": 0.010025080923212687, "darwin": 0.005712118926584214, "sagan": 0.004179075481090927, "ibn_sina": 0.015527149080130857, "einstein": 0.960630697491561, "socrates": 0.003925878097420327}, {"ibn_sina": 0.012186377714751472, "einstein": 0.96808389038311, "darwin": 0.004211499445628232, "sagan": 0.005407672330552983, "hypatia": 0.0073914117148355446, "socrates": 0.0027191484111217523}, {"einstein": 0.9626526456214559, "sagan": 0.0025400725571788196, "socrates": 0.002703893162366171, "ibn_sina": 0.0226394202673402, "darwin": 0.00782397428809651, "hypatia": 0.0016399941035624324}, {"socrates": 0.0069618592680230655, "ibn_sina": 0.004784811235363819, "hypatia": 0.002405954183771076, "einstein": 0.9706311935575636, "sagan": 0.013845310384276218, "darwin": 0.0013708713710022659}], "prompt_tokens": 345, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0031138858688066727, "einstein": 0.9783538571746611, "hypatia": 0.002140140373009538, "ibn_sina": 0.005133929866434111, "sagan": 0.009591443718706075, "socrates": 0.0016667429983825567}, "confidence": 0.9247465361875539, "choice": "einstein"}}, "usage": {"input_tokens": 345, "output_tokens": 1}, "latency_ms": 186.09654100146145, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 345, "compiled_prompt_sha256": "5b0777e33fab5b56af64797557c932888e6e56a8386aa298f23774efd807c843", "compiled_ids_sha256": "526e5ba6afca227b1070971298c7ab7274443e599e3f4513ee13c7dcebd97fa0", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.375, 21.125, 15.0, 15.875, 16.5, 14.75], "runtime_probabilities": [0.003113886108621955, 0.9783539175987244, 0.002140140626579523, 0.005133930128067732, 0.00959144439548254, 0.0016667433083057404], "decide_seconds": 0.18187284400119097, "probabilities_f64_minus_runtime_max_abs": 6.042406330841743e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.022128071403571446, "hypatia": 0.005955689622689923, "sagan": 0.019527954473822565, "socrates": 0.002813268576265473, "darwin": 0.008665477269144553, "einstein": 0.940909538654506}, "confidence": 0.8288377425360992, "choice": "einstein"}}, "usage": {"input_tokens": 345, "output_tokens": 1}, "latency_ms": 185.80341299821157, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 345, "compiled_prompt_sha256": "0d097107b6a8fe4f153aa29a3af80c9c138fd5d2816bebbbec8d60ed14143d0a", "compiled_ids_sha256": "49cab2c0d062118149ff27befc49d89cea6a55ee4887505904f2f22ea8d67763", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 15.5625, 16.75, 14.8125, 15.9375, 20.625], "runtime_probabilities": [0.022128071635961533, 0.005955689121037722, 0.01952795311808586, 0.0028132684528827667, 0.0086654769256711, 0.9409095048904419], "decide_seconds": 0.18174546499722055, "probabilities_f64_minus_runtime_max_abs": 3.3764064100161306e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.010025080923212687, "darwin": 0.005712118926584214, "sagan": 0.004179075481090927, "ibn_sina": 0.015527149080130857, "einstein": 0.960630697491561, "socrates": 0.003925878097420327}, "confidence": 0.8752372619477977, "choice": "einstein"}}, "usage": {"input_tokens": 345, "output_tokens": 1}, "latency_ms": 185.785682999267, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 345, "compiled_prompt_sha256": "08504ba6b8e244df702dd952429002075c657f9019924174db85f519dc7d4195", "compiled_ids_sha256": "ea20be630b9b05af076a10c4e9a222911b5099a3967043cea3f18ea5c9f47c59", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 15.375, 15.0625, 16.375, 20.5, 15.0], "runtime_probabilities": [0.010025082156062126, 0.005712118931114674, 0.004179075825959444, 0.01552714966237545, 0.9606307148933411, 0.003925878554582596], "decide_seconds": 0.18178455499946722, "probabilities_f64_minus_runtime_max_abs": 1.7401780016967905e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.012186377714751472, "einstein": 0.96808389038311, "darwin": 0.004211499445628232, "sagan": 0.005407672330552983, "hypatia": 0.0073914117148355446, "socrates": 0.0027191484111217523}, "confidence": 0.8946776087637641, "choice": "einstein"}}, "usage": {"input_tokens": 345, "output_tokens": 1}, "latency_ms": 186.37464900166378, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 345, "compiled_prompt_sha256": "975fb5a80fe63ff6e9b0cea5774e58f849b0b863b1fb38437460dafb81c77b98", "compiled_ids_sha256": "94209938a1eb12f2bd585124138c5f6f1aa76e3a2442bb11f86578a8721ae68e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 20.875, 15.4375, 15.6875, 16.0, 15.0], "runtime_probabilities": [0.012186378240585327, 0.968083918094635, 0.004211499821394682, 0.005407672841101885, 0.0073914118111133575, 0.002719148760661483], "decide_seconds": 0.18220053099867073, "probabilities_f64_minus_runtime_max_abs": 2.7711524963613954e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.9626526456214559, "sagan": 0.0025400725571788196, "socrates": 0.002703893162366171, "ibn_sina": 0.0226394202673402, "darwin": 0.00782397428809651, "hypatia": 0.0016399941035624324}, "confidence": 0.8872419240292067, "choice": "einstein"}}, "usage": {"input_tokens": 345, "output_tokens": 1}, "latency_ms": 186.17285099753644, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 345, "compiled_prompt_sha256": "0479b973affcdcbaddd3c7ab9b3104ed54e2f719ac405e19ef024dec2504ae60", "compiled_ids_sha256": "3668378c5efff7a76c8a6932e67084fbfebd20b93955db9a4044351891f6a111", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.75, 14.8125, 14.875, 17.0, 15.9375, 14.375], "runtime_probabilities": [0.9626526236534119, 0.0025400726590305567, 0.002703893231227994, 0.02263941988348961, 0.007823974825441837, 0.0016399942105636], "decide_seconds": 0.18192475300020305, "probabilities_f64_minus_runtime_max_abs": 2.1968044050524327e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0069618592680230655, "ibn_sina": 0.004784811235363819, "hypatia": 0.002405954183771076, "einstein": 0.9706311935575636, "sagan": 0.013845310384276218, "darwin": 0.0013708713710022659}, "confidence": 0.9040737138641108, "choice": "einstein"}}, "usage": {"input_tokens": 345, "output_tokens": 1}, "latency_ms": 185.85943300058716, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 345, "compiled_prompt_sha256": "c3c12676b722efd1aae12c93922ffc374ad0ba0d69b50bbaf70f9d7e70b298bf", "compiled_ids_sha256": "7d641b27474de1924e40dbf08e5dc53ed3bd65542e5474966c82971381d762dc", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 15.5625, 14.875, 20.875, 16.625, 14.3125], "runtime_probabilities": [0.006961858365684748, 0.004784810822457075, 0.0024059542920440435, 0.9706311821937561, 0.013845309615135193, 0.0013708713231608272], "decide_seconds": 0.1817499749995477, "probabilities_f64_minus_runtime_max_abs": 1.1363807450592844e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:03Z"} {"id": "c-einstein-01-main-03", "uid": "c028", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "prompt_sha256_by_order": ["046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3"], "request_sha256": "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "compiled_prompt_sha256_by_order": ["23475b0a8910eef4fdc6c4e53fc0b009b8dddcf3b82d3c5b95467ce8d3a7fa19", "d5c8bac853b0e1d1b21e4a6fbaef7a867d13dd471060dad8608a82330852625d", "fb7a1f72ac4b85d3efe417cd470684f4e9b8df82cdba1cd52662f6ed4e8a1694", "bcb45eeb2d4da70be6304eba0a32c51cca067e0a883fb64f92e5015e0ee15acb", "6d8188e5da56048091a1bc58aa71d86c1aa92b521c07c4e77e85b1f010de6599", "629ccc60ca658ad9707ac652d5203d19c28bfa07038925151a13f5bd43db6135"], "compiled_tokens_by_order": [358, 358, 358, 358, 358, 358], "choice": "einstein", "probs": {"darwin": 0.007027575115644449, "einstein": 0.8646646228977473, "hypatia": 0.0062018132722368065, "ibn_sina": 0.0911349807594994, "sagan": 0.02958719636622721, "socrates": 0.0013838115886448564}, "confidence": 0.7077319722639732, "seconds": 0.1920708060006291, "seconds_all_orders": 1.1524248360037745, "seconds_by_order": [0.19226105300185736, 0.19223080299707362, 0.19218755300244084, 0.19195760599905043, 0.19212191400220036, 0.19166590700115194], "argmax_stable": true, "spread": {"darwin": 0.01837176834165187, "einstein": 0.09841064167796598, "hypatia": 0.012174008971249074, "ibn_sina": 0.06918089144920948, "sagan": 0.025397486173597356, "socrates": 0.004244584932459315}, "spread_2dp": {"darwin": 0.02, "einstein": 0.09999999999999998, "hypatia": 0.01, "ibn_sina": 0.06999999999999999, "sagan": 0.03, "socrates": 0.01}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.007027575115644449, "einstein": 0.8646646228977473, "hypatia": 0.0062018132722368065, "ibn_sina": 0.0911349807594994, "sagan": 0.02958719636622721, "socrates": 0.0013838115886448564}, {"ibn_sina": 0.08261690366920249, "hypatia": 0.005984748428719187, "sagan": 0.010503560600559438, "socrates": 0.0034100068650195713, "darwin": 0.009269359696103296, "einstein": 0.888215420740396}, {"hypatia": 0.005379694375892427, "darwin": 0.003065257457685047, "sagan": 0.004189710192629855, "ibn_sina": 0.022649359312500356, "einstein": 0.9630752645757132, "socrates": 0.0016407140855789865}, {"ibn_sina": 0.02490368354558717, "einstein": 0.9345038307024691, "darwin": 0.006702735496201842, "sagan": 0.01939500824666575, "hypatia": 0.01332998122999586, "socrates": 0.00116476077908027}, {"einstein": 0.8712611610136641, "sagan": 0.010967559425148196, "socrates": 0.004294957981800193, "ibn_sina": 0.09183025076170984, "darwin": 0.020490098558930785, "hypatia": 0.0011559722587467865}, {"socrates": 0.0054093457115395845, "ibn_sina": 0.06589932146218822, "hypatia": 0.0054093457115395845, "einstein": 0.9097120719362608, "sagan": 0.011451584961193001, "darwin": 0.0021183302172789143}], "prompt_tokens": 358, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.007027575115644449, "einstein": 0.8646646228977473, "hypatia": 0.0062018132722368065, "ibn_sina": 0.0911349807594994, "sagan": 0.02958719636622721, "socrates": 0.0013838115886448564}, "confidence": 0.7077319722639732, "choice": "einstein"}}, "usage": {"input_tokens": 358, "output_tokens": 1}, "latency_ms": 192.23425300151575, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 358, "compiled_prompt_sha256": "23475b0a8910eef4fdc6c4e53fc0b009b8dddcf3b82d3c5b95467ce8d3a7fa19", "compiled_ids_sha256": "74797ebb675858e67a4e4dc5542e122d2f6918fddd42e509d7955ac569f7cd1a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.3125, 20.125, 15.1875, 17.875, 16.75, 13.6875], "runtime_probabilities": [0.007027575746178627, 0.864664614200592, 0.006201812997460365, 0.09113498032093048, 0.029587196186184883, 0.0013838117010891438], "decide_seconds": 0.18801154700122424, "probabilities_f64_minus_runtime_max_abs": 8.697155218229113e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.08261690366920249, "hypatia": 0.005984748428719187, "sagan": 0.010503560600559438, "socrates": 0.0034100068650195713, "darwin": 0.009269359696103296, "einstein": 0.888215420740396}, "confidence": 0.7474275774278079, "choice": "einstein"}}, "usage": {"input_tokens": 358, "output_tokens": 1}, "latency_ms": 192.20615299855126, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 358, "compiled_prompt_sha256": "d5c8bac853b0e1d1b21e4a6fbaef7a867d13dd471060dad8608a82330852625d", "compiled_ids_sha256": "286586acffe27ea07f1ab127988e1ced1e45101be3885588aeadb470407e222a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 15.125, 15.6875, 14.5625, 15.5625, 20.125], "runtime_probabilities": [0.0826169028878212, 0.005984748713672161, 0.010503560304641724, 0.003410007106140256, 0.009269360452890396, 0.8882154822349548], "decide_seconds": 0.18800401699991198, "probabilities_f64_minus_runtime_max_abs": 6.149455888948552e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.005379694375892427, "darwin": 0.003065257457685047, "sagan": 0.004189710192629855, "ibn_sina": 0.022649359312500356, "einstein": 0.9630752645757132, "socrates": 0.0016407140855789865}, "confidence": 0.8876342995914367, "choice": "einstein"}}, "usage": {"input_tokens": 358, "output_tokens": 1}, "latency_ms": 192.16369399873656, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 358, "compiled_prompt_sha256": "fb7a1f72ac4b85d3efe417cd470684f4e9b8df82cdba1cd52662f6ed4e8a1694", "compiled_ids_sha256": "638ec1cf3b2e2e6011bf84fb8103217a4c11dc15a7641786a107d245e93cb377", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0625, 14.5, 14.8125, 16.5, 20.25, 13.875], "runtime_probabilities": [0.0053796954452991486, 0.0030652580317109823, 0.004189710598438978, 0.022649362683296204, 0.9630753993988037, 0.0016407144721597433], "decide_seconds": 0.18803354599731392, "probabilities_f64_minus_runtime_max_abs": 1.3482309046963792e-07}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.02490368354558717, "einstein": 0.9345038307024691, "darwin": 0.006702735496201842, "sagan": 0.01939500824666575, "hypatia": 0.01332998122999586, "socrates": 0.00116476077908027}, "confidence": 0.8154285869096103, "choice": "einstein"}}, "usage": {"input_tokens": 358, "output_tokens": 1}, "latency_ms": 191.93371599976672, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 358, "compiled_prompt_sha256": "bcb45eeb2d4da70be6304eba0a32c51cca067e0a883fb64f92e5015e0ee15acb", "compiled_ids_sha256": "e8f880dbee5dcd76cf1f2c023ebcf933b782730bb236941ce9e270417bfb4055", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 20.25, 15.3125, 16.375, 16.0, 13.5625], "runtime_probabilities": [0.024903682991862297, 0.9345037937164307, 0.006702735088765621, 0.0193950068205595, 0.013329979963600636, 0.001164760789833963], "decide_seconds": 0.1878182869986631, "probabilities_f64_minus_runtime_max_abs": 3.698603845414539e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.8712611610136641, "sagan": 0.010967559425148196, "socrates": 0.004294957981800193, "ibn_sina": 0.09183025076170984, "darwin": 0.020490098558930785, "hypatia": 0.0011559722587467865}, "confidence": 0.7210966081030641, "choice": "einstein"}}, "usage": {"input_tokens": 358, "output_tokens": 1}, "latency_ms": 192.09761499951128, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 358, "compiled_prompt_sha256": "6d8188e5da56048091a1bc58aa71d86c1aa92b521c07c4e77e85b1f010de6599", "compiled_ids_sha256": "4f28680994a18df493114096b811feeca2e7fccba6ef45f94247bb442e92e9e2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.375, 16.0, 15.0625, 18.125, 16.625, 13.75], "runtime_probabilities": [0.8712611794471741, 0.010967560112476349, 0.0042949579656124115, 0.09183024615049362, 0.020490098744630814, 0.0011559722479432821], "decide_seconds": 0.18783764700128813, "probabilities_f64_minus_runtime_max_abs": 1.8433509940685155e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0054093457115395845, "ibn_sina": 0.06589932146218822, "hypatia": 0.0054093457115395845, "einstein": 0.9097120719362608, "sagan": 0.011451584961193001, "darwin": 0.0021183302172789143}, "confidence": 0.7845683451592742, "choice": "einstein"}}, "usage": {"input_tokens": 358, "output_tokens": 1}, "latency_ms": 191.64173800163553, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 358, "compiled_prompt_sha256": "629ccc60ca658ad9707ac652d5203d19c28bfa07038925151a13f5bd43db6135", "compiled_ids_sha256": "a478eb11b273267fe901ecfa66d8fd870201a40c4c2e3b7ed607f04aab81aee8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.25, 17.75, 15.25, 20.375, 16.0, 14.3125], "runtime_probabilities": [0.005409345496445894, 0.0658993273973465, 0.005409345496445894, 0.9097120761871338, 0.011451585218310356, 0.002118330216035247], "decide_seconds": 0.1875219499997911, "probabilities_f64_minus_runtime_max_abs": 5.935158281511477e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:05Z"} {"id": "c-einstein-01-main-03", "uid": "c029", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "prompt_sha256_by_order": ["1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979"], "request_sha256": "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "compiled_prompt_sha256_by_order": ["1f30f52d13f815504e7bf1165146b0c24247b90d1788ddb8d1a5b6473d203fe6", "fad5040fb67c25ac673e25a87b21d5f47d86e87e5aeec877e9176f428a184e3d", "a580632c4fd0505a5be8347d3492aeadeb88804466f558c10730641519263bba", "3cc8ae1c56a9863058534f4787fd929eef7ef7a37778b4f170d6952471538abb", "f673245e2835790db6990318e8f439fe617203b5a5956efef21eed45eea0ed61", "d7867483a63e187de51f957a47f9381866e9e5d3efd5b2458af608eae8da9eba"], "compiled_tokens_by_order": [402, 402, 402, 402, 402, 402], "choice": "einstein", "probs": {"darwin": 0.0008529962468079447, "einstein": 0.9957536309566408, "hypatia": 0.0004860226109911063, "ibn_sina": 0.0011659076338159467, "sagan": 0.0014970550353299397, "socrates": 0.0002443875164142148}, "confidence": 0.9812375349815071, "seconds": 0.21261491650147946, "seconds_all_orders": 1.2756894990088767, "seconds_by_order": [0.212391266002669, 0.21291076200213865, 0.21276431400110596, 0.21279458400022122, 0.21238662700125133, 0.21244194600149058], "argmax_stable": true, "spread": {"darwin": 0.002929927042039264, "einstein": 0.012028724934022184, "hypatia": 0.004069527484952269, "ibn_sina": 0.007346169857031454, "sagan": 0.003329269066056782, "socrates": 0.003532296162452361}, "spread_2dp": {"darwin": 0.0, "einstein": 0.020000000000000018, "hypatia": 0.0, "ibn_sina": 0.01, "sagan": 0.0, "socrates": 0.0}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0008529962468079447, "einstein": 0.9957536309566408, "hypatia": 0.0004860226109911063, "ibn_sina": 0.0011659076338159467, "sagan": 0.0014970550353299397, "socrates": 0.0002443875164142148}, {"ibn_sina": 0.0085120774908474, "hypatia": 0.0029416952660554782, "sagan": 0.0012262810298865937, "socrates": 0.0013053693613903651, "darwin": 0.0021521902322445374, "einstein": 0.9838623866195756}, {"hypatia": 0.004033937314078452, "darwin": 0.0017900488321580096, "sagan": 0.0015797125498546159, "ibn_sina": 0.004294103918458506, "einstein": 0.9870719160145959, "socrates": 0.0012302813708545005}, {"ibn_sina": 0.002961202135828447, "einstein": 0.9903865414740126, "darwin": 0.0006207016873257943, "sagan": 0.0019118957954527634, "hypatia": 0.0035718915908862885, "socrates": 0.0005477673164940456}, {"einstein": 0.9844909392845221, "sagan": 0.002153565185616418, "socrates": 0.0013904461886596502, "ibn_sina": 0.007516681079051269, "darwin": 0.0035506287293650585, "hypatia": 0.0008977395327853819}, {"socrates": 0.0037766836788665757, "ibn_sina": 0.0025956742018698443, "hypatia": 0.004555550095943376, "einstein": 0.9837249060226186, "sagan": 0.004555550095943376, "darwin": 0.0007916359047581341}], "prompt_tokens": 402, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0008529962468079447, "einstein": 0.9957536309566408, "hypatia": 0.0004860226109911063, "ibn_sina": 0.0011659076338159467, "sagan": 0.0014970550353299397, "socrates": 0.0002443875164142148}, "confidence": 0.9812375349815071, "choice": "einstein"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 212.35909600000014, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 402, "compiled_prompt_sha256": "1f30f52d13f815504e7bf1165146b0c24247b90d1788ddb8d1a5b6473d203fe6", "compiled_ids_sha256": "8e661bcb0784f987ceba2e82943447a2e43cffd2e0348e8aba9d9b56cec2f1d5", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.0625, 21.125, 13.5, 14.375, 14.625, 12.8125], "runtime_probabilities": [0.0008529962506145239, 0.9957536458969116, 0.0004860226472374052, 0.00116590759716928, 0.00149705505464226, 0.00024438754189759493], "decide_seconds": 0.20758552400002372, "probabilities_f64_minus_runtime_max_abs": 1.4940270798291522e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0085120774908474, "hypatia": 0.0029416952660554782, "sagan": 0.0012262810298865937, "socrates": 0.0013053693613903651, "darwin": 0.0021521902322445374, "einstein": 0.9838623866195756}, "confidence": 0.9420507005074771, "choice": "einstein"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 212.88563200141652, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 402, "compiled_prompt_sha256": "fad5040fb67c25ac673e25a87b21d5f47d86e87e5aeec877e9176f428a184e3d", "compiled_ids_sha256": "c34c9aa40ca503f588b1c0e1f62107facfb1588e5d86a3421a54bf52b2d80bc8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 14.9375, 14.0625, 14.125, 14.625, 20.75], "runtime_probabilities": [0.00851207785308361, 0.0029416955076158047, 0.0012262811651453376, 0.0013053694274276495, 0.002152190310880542, 0.9838624596595764], "decide_seconds": 0.2082673189979687, "probabilities_f64_minus_runtime_max_abs": 7.304000082530848e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.004033937314078452, "darwin": 0.0017900488321580096, "sagan": 0.0015797125498546159, "ibn_sina": 0.004294103918458506, "einstein": 0.9870719160145959, "socrates": 0.0012302813708545005}, "confidence": 0.9507496317300203, "choice": "einstein"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 212.73892399767647, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 402, "compiled_prompt_sha256": "a580632c4fd0505a5be8347d3492aeadeb88804466f558c10730641519263bba", "compiled_ids_sha256": "686f4e052bae09a65511588e158762eed8dc752bf7ce44cfec785dd6221a492a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0, 14.1875, 14.0625, 15.0625, 20.5, 13.8125], "runtime_probabilities": [0.004033937584608793, 0.0017900487873703241, 0.0015797126106917858, 0.004294103942811489, 0.9870719313621521, 0.0012302814284339547], "decide_seconds": 0.20812214900070103, "probabilities_f64_minus_runtime_max_abs": 1.5347556225897563e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.002961202135828447, "einstein": 0.9903865414740126, "darwin": 0.0006207016873257943, "sagan": 0.0019118957954527634, "hypatia": 0.0035718915908862885, "socrates": 0.0005477673164940456}, "confidence": 0.9622722115117335, "choice": "einstein"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 212.76834400123334, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 402, "compiled_prompt_sha256": "3cc8ae1c56a9863058534f4787fd929eef7ef7a37778b4f170d6952471538abb", "compiled_ids_sha256": "66e0ef5d15d470cf399bedf004fc077b33cf6bb3d11369cf2a24cead667e7e70", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.4375, 21.25, 13.875, 15.0, 15.625, 13.75], "runtime_probabilities": [0.002961202058941126, 0.9903864860534668, 0.0006207015831023455, 0.0019118956988677382, 0.0035718914587050676, 0.000547767267562449], "decide_seconds": 0.2078358719991229, "probabilities_f64_minus_runtime_max_abs": 5.5420545819195866e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.9844909392845221, "sagan": 0.002153565185616418, "socrates": 0.0013904461886596502, "ibn_sina": 0.007516681079051269, "darwin": 0.0035506287293650585, "hypatia": 0.0008977395327853819}, "confidence": 0.943716631884768, "choice": "einstein"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 212.36075700289803, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 402, "compiled_prompt_sha256": "f673245e2835790db6990318e8f439fe617203b5a5956efef21eed45eea0ed61", "compiled_ids_sha256": "c02af61b97cc7e31a8f7047b29fa231950113da8e17d708229ee8d4e37624dab", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.875, 14.75, 14.3125, 16.0, 15.25, 13.875], "runtime_probabilities": [0.9844909310340881, 0.002153565175831318, 0.0013904462102800608, 0.007516680750995874, 0.0035506286658346653, 0.0008977395482361317], "decide_seconds": 0.207723701998475, "probabilities_f64_minus_runtime_max_abs": 8.250434002654572e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0037766836788665757, "ibn_sina": 0.0025956742018698443, "hypatia": 0.004555550095943376, "einstein": 0.9837249060226186, "sagan": 0.004555550095943376, "darwin": 0.0007916359047581341}, "confidence": 0.9400359379696204, "choice": "einstein"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 212.41700700193178, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 402, "compiled_prompt_sha256": "d7867483a63e187de51f957a47f9381866e9e5d3efd5b2458af608eae8da9eba", "compiled_ids_sha256": "8ad8767723780577622e71e68ef170e49ca04750277a35575b3d6c85464e7a7a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.3125, 14.9375, 15.5, 20.875, 15.5, 13.75], "runtime_probabilities": [0.0037766839377582073, 0.0025956742465496063, 0.004555550403892994, 0.9837248921394348, 0.004555550403892994, 0.0007916359463706613], "decide_seconds": 0.2076756429996749, "probabilities_f64_minus_runtime_max_abs": 1.3883183824781042e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:06Z"} {"id": "c-einstein-01-starter-01", "uid": "c030", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "prompt_sha256_by_order": ["08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa"], "request_sha256": "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "compiled_prompt_sha256_by_order": ["53cefe6d23d0ac7f3c46766dc20453dc7cb5499f889b97160ea8df183525bf2a", "fe7273a07d6d72bbbd5831f96a112ff514ec70058bfdca6741d13387e4ab5ef0", "15a6f6b712c4636d2c7ce9542482f92a3e9b34461b62a779ae89a82341c9c585", "ead9536380dfdb2c9e943e386ca4e01bede87b5bd9e36e3ea213aaf5177298bc", "37131f7cc41a227893b98bf82977596196305071846c5d2130db0a9fcc58a512", "e6ed2d8a7d165361f474e7cd117bd802f4b52681f976b8a960853337c4852e81"], "compiled_tokens_by_order": [305, 305, 305, 305, 305, 305], "choice": "einstein", "probs": {"darwin": 0.0006653237158637849, "einstein": 0.9972668967659153, "hypatia": 0.0002159989859472034, "ibn_sina": 0.0004867612288347998, "sagan": 0.0010304735295296771, "socrates": 0.00033454577390927694}, "confidence": 0.9872209732578074, "seconds": 0.17590496200015573, "seconds_all_orders": 1.0554297720009345, "seconds_by_order": [0.17596392999985255, 0.17567080200024066, 0.17593629999828408, 0.17588080999848899, 0.17559522300143726, 0.17638270700263092], "argmax_stable": true, "spread": {"darwin": 0.001352856583582768, "einstein": 0.007397924065917327, "hypatia": 0.0024005675494049314, "ibn_sina": 0.004392884318876625, "sagan": 0.0005767655909610064, "socrates": 0.0022886647576097503}, "spread_2dp": {"darwin": 0.0, "einstein": 0.010000000000000009, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.0, "socrates": 0.0}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0006653237158637849, "einstein": 0.9972668967659153, "hypatia": 0.0002159989859472034, "ibn_sina": 0.0004867612288347998, "sagan": 0.0010304735295296771, "socrates": 0.00033454577390927694}, {"ibn_sina": 0.004879645547711425, "hypatia": 0.0015841889668986998, "sagan": 0.0005827905518526582, "socrates": 0.0013980418563967927, "darwin": 0.0016863603771424742, "einstein": 0.989868972699998}, {"hypatia": 0.002616566535352135, "darwin": 0.00158702782686913, "sagan": 0.0005838349100721439, "ibn_sina": 0.0014005471415275686, "einstein": 0.9916428137387157, "socrates": 0.0021692098474633257}, {"ibn_sina": 0.0020455552776475305, "einstein": 0.9954245026945666, "darwin": 0.0003339277195524703, "sagan": 0.0005505537341025449, "hypatia": 0.001094906840028177, "socrates": 0.0005505537341025449}, {"einstein": 0.9945658766643901, "sagan": 0.0005167512492499488, "socrates": 0.0008519787762992541, "ibn_sina": 0.0026242793816775288, "darwin": 0.001164516916520643, "hypatia": 0.00027659701186250036}, {"socrates": 0.002623210531519027, "ibn_sina": 0.0012391169158875676, "hypatia": 0.0005498547983010719, "einstein": 0.9941607971205216, "sagan": 0.0010935168402109552, "darwin": 0.0003335037935597061}], "prompt_tokens": 305, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0006653237158637849, "einstein": 0.9972668967659153, "hypatia": 0.0002159989859472034, "ibn_sina": 0.0004867612288347998, "sagan": 0.0010304735295296771, "socrates": 0.00033454577390927694}, "confidence": 0.9872209732578074, "choice": "einstein"}}, "usage": {"input_tokens": 305, "output_tokens": 1}, "latency_ms": 175.9368600032758, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 305, "compiled_prompt_sha256": "53cefe6d23d0ac7f3c46766dc20453dc7cb5499f889b97160ea8df183525bf2a", "compiled_ids_sha256": "f3e5ac56a91af44389f871a8caca89d7e9b0d855b71d7a5f4fe583f155b126f2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [13.8125, 21.125, 12.6875, 13.5, 14.25, 13.125], "runtime_probabilities": [0.0006653238087892532, 0.9972670078277588, 0.00021599899628199637, 0.0004867613024543971, 0.0010304736206308007, 0.000334545795340091], "decide_seconds": 0.17217109899866045, "probabilities_f64_minus_runtime_max_abs": 1.1106184349962689e-07}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.004879645547711425, "hypatia": 0.0015841889668986998, "sagan": 0.0005827905518526582, "socrates": 0.0013980418563967927, "darwin": 0.0016863603771424742, "einstein": 0.989868972699998}, "confidence": 0.960617632056358, "choice": "einstein"}}, "usage": {"input_tokens": 305, "output_tokens": 1}, "latency_ms": 175.648402000661, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 305, "compiled_prompt_sha256": "fe7273a07d6d72bbbd5831f96a112ff514ec70058bfdca6741d13387e4ab5ef0", "compiled_ids_sha256": "eb66ac5f446e0c30890d36c02b689c2df74bf26f69982ae2248faa56edd4ddbb", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.3125, 14.1875, 13.1875, 14.0625, 14.25, 20.625], "runtime_probabilities": [0.00487964553758502, 0.0015841890126466751, 0.0005827905260957778, 0.001398041844367981, 0.001686360570602119, 0.9898689985275269], "decide_seconds": 0.17199893100041663, "probabilities_f64_minus_runtime_max_abs": 2.5827528893351825e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.002616566535352135, "darwin": 0.00158702782686913, "sagan": 0.0005838349100721439, "ibn_sina": 0.0014005471415275686, "einstein": 0.9916428137387157, "socrates": 0.0021692098474633257}, "confidence": 0.9659751027443627, "choice": "einstein"}}, "usage": {"input_tokens": 305, "output_tokens": 1}, "latency_ms": 175.91329000060796, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 305, "compiled_prompt_sha256": "15a6f6b712c4636d2c7ce9542482f92a3e9b34461b62a779ae89a82341c9c585", "compiled_ids_sha256": "788547cc217fe137c8f0e6258247a249180ce82cf2d36c66eef85ce4af7a6301", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.5625, 14.0625, 13.0625, 13.9375, 20.5, 14.375], "runtime_probabilities": [0.0026165666058659554, 0.0015870279166847467, 0.0005838348879478872, 0.0014005472185090184, 0.9916428327560425, 0.002169209998100996], "decide_seconds": 0.17217837999851326, "probabilities_f64_minus_runtime_max_abs": 1.901732682441093e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0020455552776475305, "einstein": 0.9954245026945666, "darwin": 0.0003339277195524703, "sagan": 0.0005505537341025449, "hypatia": 0.001094906840028177, "socrates": 0.0005505537341025449}, "confidence": 0.980113597030822, "choice": "einstein"}}, "usage": {"input_tokens": 305, "output_tokens": 1}, "latency_ms": 175.85893099749228, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 305, "compiled_prompt_sha256": "ead9536380dfdb2c9e943e386ca4e01bede87b5bd9e36e3ea213aaf5177298bc", "compiled_ids_sha256": "3b657d2333b7c8103126966e114c0abd541610cfa1782d2d1c3ab51889d46823", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.9375, 21.125, 13.125, 13.625, 14.3125, 13.625], "runtime_probabilities": [0.0020455550402402878, 0.9954244494438171, 0.0003339276881888509, 0.0005505536682903767, 0.0010949067072942853, 0.0005505536682903767], "decide_seconds": 0.1720710199988389, "probabilities_f64_minus_runtime_max_abs": 5.325074947659658e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.9945658766643901, "sagan": 0.0005167512492499488, "socrates": 0.0008519787762992541, "ibn_sina": 0.0026242793816775288, "darwin": 0.001164516916520643, "hypatia": 0.00027659701186250036}, "confidence": 0.9770723916476761, "choice": "einstein"}}, "usage": {"input_tokens": 305, "output_tokens": 1}, "latency_ms": 175.57312300050398, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 305, "compiled_prompt_sha256": "37131f7cc41a227893b98bf82977596196305071846c5d2130db0a9fcc58a512", "compiled_ids_sha256": "fec24d9eede9f24754906876e38c33cb9abdb43055f366e6aecd3d7ce955483a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.75, 13.1875, 13.6875, 14.8125, 14.0, 12.5625], "runtime_probabilities": [0.9945658445358276, 0.000516751257237047, 0.0008519787224940956, 0.0026242793537676334, 0.0011645168997347355, 0.00027659701299853623], "decide_seconds": 0.17196747100024368, "probabilities_f64_minus_runtime_max_abs": 3.2128562477140576e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.002623210531519027, "ibn_sina": 0.0012391169158875676, "hypatia": 0.0005498547983010719, "einstein": 0.9941607971205216, "sagan": 0.0010935168402109552, "darwin": 0.0003335037935597061}, "confidence": 0.9754655967208744, "choice": "einstein"}}, "usage": {"input_tokens": 305, "output_tokens": 1}, "latency_ms": 176.36084699915955, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 305, "compiled_prompt_sha256": "e6ed2d8a7d165361f474e7cd117bd802f4b52681f976b8a960853337c4852e81", "compiled_ids_sha256": "1fcf1f17b481dfc69aef5d0a87736aa7724c202eab5cc63d8fc32403a2b73ce1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.9375, 14.1875, 13.375, 20.875, 14.0625, 12.875], "runtime_probabilities": [0.0026232104282826185, 0.0012391168856993318, 0.0005498547689057887, 0.9941607117652893, 0.0010935167083516717, 0.0003335037617944181], "decide_seconds": 0.17274447500312817, "probabilities_f64_minus_runtime_max_abs": 8.535523232655606e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:07Z"} {"id": "c-einstein-01-starter-02", "uid": "c031", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "prompt_sha256_by_order": ["06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7"], "request_sha256": "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "compiled_prompt_sha256_by_order": ["2c70462d4e85f4e64653498d9e754bf182bcbee31c8f1c5ece9f76079ffaacbf", "6ddbb2d9bc5a75bbb81da84886f84506b5ce64e9a7694668a9929a118db819ff", "0ae9b2b659ab144b74d101b700b2ff419a77540d4adc4d863325bc788325f9fa", "67887bf11c2ae2b8bce633d7ccbc05227fb3b22aed8d550200baa7aa4f1a89cb", "0444e28a162e609cf2a8b55f72729e5a229ab85fded73476d06e85590d5f61ef", "66e1cca3e42e537271045a99ce2635505d964d651ca26184d9f91ebefb77aedf"], "compiled_tokens_by_order": [267, 267, 267, 267, 267, 267], "choice": "sagan", "probs": {"darwin": 0.020399310664232152, "einstein": 0.3615864333089703, "hypatia": 0.010918964166939573, "ibn_sina": 0.008503697843541151, "sagan": 0.5961552437930927, "socrates": 0.002436350223224194}, "confidence": 0.5199625991544063, "seconds": 0.17172022799968545, "seconds_all_orders": 1.0303213679981127, "seconds_by_order": [0.17189247200076352, 0.17126095599815017, 0.1712444670010882, 0.17151137399923755, 0.17203906099894084, 0.17237303799993242], "argmax_stable": false, "spread": {"darwin": 0.021050656750766457, "einstein": 0.6102922947832086, "hypatia": 0.02393274227629391, "ibn_sina": 0.012430923140507108, "sagan": 0.5844777062576076, "socrates": 0.009383114704698924}, "spread_2dp": {"darwin": 0.02, "einstein": 0.6100000000000001, "hypatia": 0.02, "ibn_sina": 0.01, "sagan": 0.5800000000000001, "socrates": 0.01}, "choices_by_order": ["sagan", "einstein", "einstein", "sagan", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.020399310664232152, "einstein": 0.3615864333089703, "hypatia": 0.010918964166939573, "ibn_sina": 0.008503697843541151, "sagan": 0.5961552437930927, "socrates": 0.002436350223224194}, {"ibn_sina": 0.017486221302742855, "hypatia": 0.008259906077243138, "sagan": 0.21302578541618925, "socrates": 0.005676944890709605, "darwin": 0.012018092427941472, "einstein": 0.7435330498851737}, {"hypatia": 0.012247336498540215, "darwin": 0.004505543306921584, "sagan": 0.058428774624308726, "ibn_sina": 0.008960342795300662, "einstein": 0.9139798128922565, "socrates": 0.0018781898826723171}, {"ibn_sina": 0.016094848821113495, "einstein": 0.3036875181090479, "darwin": 0.011061817038512823, "sagan": 0.6429064808819164, "hypatia": 0.024928189506076438, "socrates": 0.0013211456433329602}, {"einstein": 0.7048160412460645, "sagan": 0.25928733138227045, "socrates": 0.0057284032923290065, "ibn_sina": 0.005055298162235747, "darwin": 0.02411747868731792, "hypatia": 0.0009954472297825246}, {"socrates": 0.010704260348031885, "ibn_sina": 0.005056334560242156, "hypatia": 0.006099103697756821, "einstein": 0.6622491373783131, "sagan": 0.3128243420791045, "darwin": 0.0030668219365514626}], "prompt_tokens": 267, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.020399310664232152, "einstein": 0.3615864333089703, "hypatia": 0.010918964166939573, "ibn_sina": 0.008503697843541151, "sagan": 0.5961552437930927, "socrates": 0.002436350223224194}, "confidence": 0.5199625991544063, "choice": "sagan"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 171.86566199961817, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 267, "compiled_prompt_sha256": "2c70462d4e85f4e64653498d9e754bf182bcbee31c8f1c5ece9f76079ffaacbf", "compiled_ids_sha256": "fc5d36802beb2fda14c96e74d28c496080c202d461eda3e170bc4d68d9dc027a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 19.125, 15.625, 15.375, 19.625, 14.125], "runtime_probabilities": [0.020399311557412148, 0.36158645153045654, 0.010918964631855488, 0.00850369781255722, 0.5961552858352661, 0.0024363503325730562], "decide_seconds": 0.16840765899905819, "probabilities_f64_minus_runtime_max_abs": 4.2042173409306827e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.017486221302742855, "hypatia": 0.008259906077243138, "sagan": 0.21302578541618925, "socrates": 0.005676944890709605, "darwin": 0.012018092427941472, "einstein": 0.7435330498851737}, "confidence": 0.585537594932743, "choice": "einstein"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 171.23833599907812, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 267, "compiled_prompt_sha256": "6ddbb2d9bc5a75bbb81da84886f84506b5ce64e9a7694668a9929a118db819ff", "compiled_ids_sha256": "f1baa560653c22c4c9c528290a1b434b24533cb07f24573cdb7a6d20975dd839", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 15.5, 18.75, 15.125, 15.875, 20.0], "runtime_probabilities": [0.017486222088336945, 0.008259906433522701, 0.2130257785320282, 0.005676944274455309, 0.012018091976642609, 0.7435330152511597], "decide_seconds": 0.16792654300297727, "probabilities_f64_minus_runtime_max_abs": 3.4634013990064716e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.012247336498540215, "darwin": 0.004505543306921584, "sagan": 0.058428774624308726, "ibn_sina": 0.008960342795300662, "einstein": 0.9139798128922565, "socrates": 0.0018781898826723171}, "confidence": 0.7876717528948955, "choice": "einstein"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 171.22197699791286, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 267, "compiled_prompt_sha256": "0ae9b2b659ab144b74d101b700b2ff419a77540d4adc4d863325bc788325f9fa", "compiled_ids_sha256": "f209506a405f4b1f64945fe9b066bb183c8cd249ff473cd5d22c9b503b415039", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 14.9375, 17.5, 15.625, 20.25, 14.0625], "runtime_probabilities": [0.012247336097061634, 0.004505543038249016, 0.058428771793842316, 0.008960342966020107, 0.9139797687530518, 0.0018781897379085422], "decide_seconds": 0.1678603030013619, "probabilities_f64_minus_runtime_max_abs": 4.4139204735671456e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.016094848821113495, "einstein": 0.3036875181090479, "darwin": 0.011061817038512823, "sagan": 0.6429064808819164, "hypatia": 0.024928189506076438, "socrates": 0.0013211456433329602}, "confidence": 0.5183498388604797, "choice": "sagan"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 171.4885649998905, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 267, "compiled_prompt_sha256": "67887bf11c2ae2b8bce633d7ccbc05227fb3b22aed8d550200baa7aa4f1a89cb", "compiled_ids_sha256": "b7c191be89fc6f200713caae6420ff12ad72d98d1ef4b53aa41a03e772dc232f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 18.875, 15.5625, 19.625, 16.375, 13.4375], "runtime_probabilities": [0.01609485037624836, 0.30368754267692566, 0.011061817407608032, 0.6429064869880676, 0.02492818981409073, 0.0013211456825956702], "decide_seconds": 0.16795530200033681, "probabilities_f64_minus_runtime_max_abs": 2.4567877732195598e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.7048160412460645, "sagan": 0.25928733138227045, "socrates": 0.0057284032923290065, "ibn_sina": 0.005055298162235747, "darwin": 0.02411747868731792, "hypatia": 0.0009954472297825246}, "confidence": 0.5816607209768831, "choice": "einstein"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 172.01661099898047, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 267, "compiled_prompt_sha256": "0444e28a162e609cf2a8b55f72729e5a229ab85fded73476d06e85590d5f61ef", "compiled_ids_sha256": "c9b5291c54ba363da61bbb5776949cccdab213fb6e3656843661917c82072d4d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.75, 18.75, 14.9375, 14.8125, 16.375, 13.1875], "runtime_probabilities": [0.7048159837722778, 0.25928735733032227, 0.0057284035719931126, 0.005055297631770372, 0.024117477238178253, 0.0009954471606761217], "decide_seconds": 0.16865668700120295, "probabilities_f64_minus_runtime_max_abs": 5.7473786618800204e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.010704260348031885, "ibn_sina": 0.005056334560242156, "hypatia": 0.006099103697756821, "einstein": 0.6622491373783131, "sagan": 0.3128243420791045, "darwin": 0.0030668219365514626}, "confidence": 0.5754954030248024, "choice": "einstein"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 172.3501779997605, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 267, "compiled_prompt_sha256": "66e1cca3e42e537271045a99ce2635505d964d651ca26184d9f91ebefb77aedf", "compiled_ids_sha256": "6aef6b34bf97625f7f131a5d5f75df2857a47756da3f3470ae57d51c8fde5754", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5, 14.75, 14.9375, 19.625, 18.875, 14.25], "runtime_probabilities": [0.010704259388148785, 0.005056333728134632, 0.006099103018641472, 0.6622490882873535, 0.3128243386745453, 0.0030668219551444054], "decide_seconds": 0.16902238400143688, "probabilities_f64_minus_runtime_max_abs": 4.909095963068211e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:08Z"} {"id": "c-einstein-01-starter-02", "uid": "c032", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "prompt_sha256_by_order": ["8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db"], "request_sha256": "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "compiled_prompt_sha256_by_order": ["370a32f6cf6865231ae43ab9a2c0130f1bd393f6ff9c93f7322c5eec6f69a7f6", "e52019829e848e766454e6a6ea48668b93f4bcca340e9c914614ef8f170050d4", "3b3121397ccf675eaf9eea7a8f22638372036d0a56b6ccaf93cff2422d317294", "4e63cd420f0e8ae84d5b86bdea3c661c9b72ffadd05ea27f762b7d6f36a6024c", "f60a7fd29e7b7e6df5f25948d1fc5cb160b82afd34a998f4d5847d53c378da07", "b076d723c9254fe45a7e5a59c0a520f7484fd0ab3cbd9891413450ae9e75c36d"], "compiled_tokens_by_order": [294, 294, 294, 294, 294, 294], "choice": "sagan", "probs": {"darwin": 0.0527853279826753, "einstein": 0.0527853279826753, "hypatia": 0.03627878999995784, "ibn_sina": 0.014206978293009476, "sagan": 0.8257014545191268, "socrates": 0.01824212122255519}, "confidence": 0.5967782073764261, "seconds": 0.1743044780002189, "seconds_all_orders": 1.0458268680013134, "seconds_by_order": [0.17436968200127012, 0.17436343299777946, 0.17426868400070816, 0.1744311820002622, 0.17424099300114904, 0.17415289400014444], "argmax_stable": true, "spread": {"darwin": 0.04928508772800157, "einstein": 0.11940762524100686, "hypatia": 0.015156000666415665, "ibn_sina": 0.023087335562082752, "sagan": 0.184921915514201, "socrates": 0.02929678120084949}, "spread_2dp": {"darwin": 0.049999999999999996, "einstein": 0.12000000000000001, "hypatia": 0.02, "ibn_sina": 0.019999999999999997, "sagan": 0.19000000000000006, "socrates": 0.03}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0527853279826753, "einstein": 0.0527853279826753, "hypatia": 0.03627878999995784, "ibn_sina": 0.014206978293009476, "sagan": 0.8257014545191268, "socrates": 0.01824212122255519}, {"ibn_sina": 0.03135913871620654, "hypatia": 0.02442252178862646, "sagan": 0.7137307074042132, "socrates": 0.03135913871620654, "darwin": 0.058586584135109164, "einstein": 0.1405419092396381}, {"hypatia": 0.028650625953323343, "darwin": 0.009301496407107594, "sagan": 0.8372940233699362, "ibn_sina": 0.013533597415443596, "einstein": 0.10000051052305559, "socrates": 0.01121974633113375}, {"ibn_sina": 0.02113428399863125, "einstein": 0.02113428399863125, "darwin": 0.01645939692778748, "sagan": 0.8986526229184142, "hypatia": 0.0348445435695607, "socrates": 0.0077748685869750206}, {"einstein": 0.05368725197840523, "sagan": 0.8398099195718461, "socrates": 0.028736715187220166, "ibn_sina": 0.013574263090087553, "darwin": 0.041811673881733825, "hypatia": 0.022380176290707}, {"socrates": 0.03707164978782451, "ibn_sina": 0.008271803154123788, "hypatia": 0.021122789333542173, "einstein": 0.07848068321668375, "sagan": 0.8437468601148482, "darwin": 0.011306214392977472}], "prompt_tokens": 294, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0527853279826753, "einstein": 0.0527853279826753, "hypatia": 0.03627878999995784, "ibn_sina": 0.014206978293009476, "sagan": 0.8257014545191268, "socrates": 0.01824212122255519}, "confidence": 0.5967782073764261, "choice": "sagan"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 174.34454200338223, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 294, "compiled_prompt_sha256": "370a32f6cf6865231ae43ab9a2c0130f1bd393f6ff9c93f7322c5eec6f69a7f6", "compiled_ids_sha256": "45bd3725a496f6775e0e3d110492d7c95aeb7acbd0c5e55b8758c122228a0ab4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 16.875, 16.5, 15.5625, 19.625, 15.8125], "runtime_probabilities": [0.052785325795412064, 0.052785325795412064, 0.03627878800034523, 0.014206977561116219, 0.8257014155387878, 0.01824212074279785], "decide_seconds": 0.17066697200061753, "probabilities_f64_minus_runtime_max_abs": 3.898033895666231e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.03135913871620654, "hypatia": 0.02442252178862646, "sagan": 0.7137307074042132, "socrates": 0.03135913871620654, "darwin": 0.058586584135109164, "einstein": 0.1405419092396381}, "confidence": 0.44718180304882205, "choice": "sagan"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 174.34135299845366, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 294, "compiled_prompt_sha256": "e52019829e848e766454e6a6ea48668b93f4bcca340e9c914614ef8f170050d4", "compiled_ids_sha256": "0b825118dfe92a780f2d6b3cb156be8e2c51d191e0e73f63c14f0973f77e565a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 15.875, 19.25, 16.125, 16.75, 17.625], "runtime_probabilities": [0.03135913982987404, 0.024422520771622658, 0.7137306928634644, 0.03135913982987404, 0.05858658254146576, 0.1405418962240219], "decide_seconds": 0.17071581100026378, "probabilities_f64_minus_runtime_max_abs": 1.4540748827229777e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.028650625953323343, "darwin": 0.009301496407107594, "sagan": 0.8372940233699362, "ibn_sina": 0.013533597415443596, "einstein": 0.10000051052305559, "socrates": 0.01121974633113375}, "confidence": 0.6468025282842047, "choice": "sagan"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 174.24725399905583, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 294, "compiled_prompt_sha256": "3b3121397ccf675eaf9eea7a8f22638372036d0a56b6ccaf93cff2422d317294", "compiled_ids_sha256": "943ef61d197eea000fa0604661bd388bd860b1b6552a1a21d94cb1c85af2426c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 15.25, 19.75, 15.625, 17.625, 15.4375], "runtime_probabilities": [0.028650622814893723, 0.009301496669650078, 0.8372939825057983, 0.013533596880733967, 0.10000050812959671, 0.011219745501875877], "decide_seconds": 0.1706666509999195, "probabilities_f64_minus_runtime_max_abs": 4.086413785131526e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.02113428399863125, "einstein": 0.02113428399863125, "darwin": 0.01645939692778748, "sagan": 0.8986526229184142, "hypatia": 0.0348445435695607, "socrates": 0.0077748685869750206}, "confidence": 0.7313373578240709, "choice": "sagan"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 174.40926199924434, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 294, "compiled_prompt_sha256": "4e63cd420f0e8ae84d5b86bdea3c661c9b72ffadd05ea27f762b7d6f36a6024c", "compiled_ids_sha256": "8b018fb09a854ef86715772ca5aec50aa39589b6d733ca573546abe6e0bb32fd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 16.125, 15.875, 19.875, 16.625, 15.125], "runtime_probabilities": [0.021134285256266594, 0.021134285256266594, 0.016459397971630096, 0.8986526727676392, 0.034844543784856796, 0.007774868980050087], "decide_seconds": 0.17076026100039599, "probabilities_f64_minus_runtime_max_abs": 4.984922497008881e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.05368725197840523, "sagan": 0.8398099195718461, "socrates": 0.028736715187220166, "ibn_sina": 0.013574263090087553, "darwin": 0.041811673881733825, "hypatia": 0.022380176290707}, "confidence": 0.619500796101726, "choice": "sagan"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 174.21955299869296, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 294, "compiled_prompt_sha256": "f60a7fd29e7b7e6df5f25948d1fc5cb160b82afd34a998f4d5847d53c378da07", "compiled_ids_sha256": "545c034c3c21ebbfc4845d601901f40ca7982940480a018d8a8d0660d31376a4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 19.75, 16.375, 15.625, 16.75, 16.125], "runtime_probabilities": [0.05368725210428238, 0.8398099541664124, 0.028736716136336327, 0.013574263080954552, 0.04181167483329773, 0.022380176931619644], "decide_seconds": 0.17064582100283587, "probabilities_f64_minus_runtime_max_abs": 3.459456621168755e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.03707164978782451, "ibn_sina": 0.008271803154123788, "hypatia": 0.021122789333542173, "einstein": 0.07848068321668375, "sagan": 0.8437468601148482, "darwin": 0.011306214392977472}, "confidence": 0.6444563446489293, "choice": "sagan"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 174.1321350018552, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 294, "compiled_prompt_sha256": "b076d723c9254fe45a7e5a59c0a520f7484fd0ab3cbd9891413450ae9e75c36d", "compiled_ids_sha256": "75b3c95af00e5cc43dd13c6f18112755b3bfdc54a67f447b092a77fde8f7f9c9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 15.0, 15.9375, 17.25, 19.625, 15.3125], "runtime_probabilities": [0.03707164525985718, 0.00827180314809084, 0.02112278901040554, 0.07848067581653595, 0.8437467813491821, 0.011306213214993477], "decide_seconds": 0.1705769220025104, "probabilities_f64_minus_runtime_max_abs": 7.876566610054425e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:09Z"} {"id": "c-einstein-01-starter-03", "uid": "c033", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "prompt_sha256_by_order": ["ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c"], "request_sha256": "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "compiled_prompt_sha256_by_order": ["32dea34ab78631d1128469f41a9a9b72979f369974ba4b1ff7cdc18dca902d44", "b2dc88d839a341d8e5dbb739cd7f24a26c813ab1a286765ea240e741b32f962d", "b429c3380c2f0253b5f360120a373ebdfca865dff32f4f83aa8149409771f11a", "32569ed4b1860f0ccaa5b47061cf534a8b0cec2159b6ea6dcf44768ba0300d0f", "d95158421940fd1c60b8d6778bc47021fd09a020413870cd075808a0690e9efb", "7e24416d18a1e23b747b58650780df0865e19627020d3b2ca042eb06a7df934c"], "compiled_tokens_by_order": [387, 387, 387, 387, 387, 387], "choice": "einstein", "probs": {"darwin": 0.004663134723531164, "einstein": 0.9459482934182293, "hypatia": 0.00363165297425353, "ibn_sina": 0.011907731671536362, "sagan": 0.028565163314038113, "socrates": 0.005284023898411542}, "confidence": 0.8437141709717673, "seconds": 0.21062840833292285, "seconds_all_orders": 1.263770449997537, "seconds_by_order": [0.21066272000098252, 0.21064439000110724, 0.21083275899945875, 0.21044072099903133, 0.210720818999107, 0.21046904099785024], "argmax_stable": true, "spread": {"darwin": 0.004025498721424942, "einstein": 0.03205507364108984, "hypatia": 0.016061845910442147, "ibn_sina": 0.007403549982580429, "sagan": 0.022391301859290117, "socrates": 0.003911994421026473}, "spread_2dp": {"darwin": 0.0, "einstein": 0.040000000000000036, "hypatia": 0.02, "ibn_sina": 0.01, "sagan": 0.019999999999999997, "socrates": 0.01}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.004663134723531164, "einstein": 0.9459482934182293, "hypatia": 0.00363165297425353, "ibn_sina": 0.011907731671536362, "sagan": 0.028565163314038113, "socrates": 0.005284023898411542}, {"ibn_sina": 0.010750864697647565, "hypatia": 0.00421009910723975, "sagan": 0.006941279949861796, "socrates": 0.006125658055725591, "darwin": 0.00421009910723975, "einstein": 0.9677619990822857}, {"hypatia": 0.0051130518765475684, "darwin": 0.0020022998843463933, "sagan": 0.008429997387057181, "ibn_sina": 0.00656528856633003, "einstein": 0.9743752165490657, "socrates": 0.003514145736653205}, {"ibn_sina": 0.009844451990970683, "einstein": 0.943322338736569, "darwin": 0.0036215714970773876, "sagan": 0.01957803054194739, "hypatia": 0.017277551311975177, "socrates": 0.006356055921460349}, {"einstein": 0.9753774123776588, "sagan": 0.006173861454747995, "socrates": 0.003986144550628004, "ibn_sina": 0.008438668079900008, "darwin": 0.004808208135532101, "hypatia": 0.00121570540153303}, {"socrates": 0.007426140157679677, "ibn_sina": 0.004504181688955933, "hypatia": 0.0024109147251396316, "einstein": 0.9726324187769506, "sagan": 0.012243635237166888, "darwin": 0.0007827094141071589}], "prompt_tokens": 387, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.004663134723531164, "einstein": 0.9459482934182293, "hypatia": 0.00363165297425353, "ibn_sina": 0.011907731671536362, "sagan": 0.028565163314038113, "socrates": 0.005284023898411542}, "confidence": 0.8437141709717673, "choice": "einstein"}}, "usage": {"input_tokens": 387, "output_tokens": 1}, "latency_ms": 210.63383000000613, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 387, "compiled_prompt_sha256": "32dea34ab78631d1128469f41a9a9b72979f369974ba4b1ff7cdc18dca902d44", "compiled_ids_sha256": "4844b40c337845fad28c4f6e260cd19b165cd0fb8800c232d4d2050f7333a62b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.1875, 20.5, 14.9375, 16.125, 17.0, 15.3125], "runtime_probabilities": [0.0046631344594061375, 0.9459483027458191, 0.003631653031334281, 0.011907732114195824, 0.028565164655447006, 0.005284023936837912], "decide_seconds": 0.20607251500041457, "probabilities_f64_minus_runtime_max_abs": 9.327589811647385e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.010750864697647565, "hypatia": 0.00421009910723975, "sagan": 0.006941279949861796, "socrates": 0.006125658055725591, "darwin": 0.00421009910723975, "einstein": 0.9677619990822857}, "confidence": 0.892721915779305, "choice": "einstein"}}, "usage": {"input_tokens": 387, "output_tokens": 1}, "latency_ms": 210.61814999848139, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 387, "compiled_prompt_sha256": "b2dc88d839a341d8e5dbb739cd7f24a26c813ab1a286765ea240e741b32f962d", "compiled_ids_sha256": "81c845b35e69d0dcc40433722e5747ef371bb7881b11bc6acb0d5dd0398ef11c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 15.1875, 15.6875, 15.5625, 15.1875, 20.625], "runtime_probabilities": [0.010750866495072842, 0.0042100995779037476, 0.006941279862076044, 0.00612565828487277, 0.0042100995779037476, 0.9677620530128479], "decide_seconds": 0.20610603499881108, "probabilities_f64_minus_runtime_max_abs": 5.393056223823578e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0051130518765475684, "darwin": 0.0020022998843463933, "sagan": 0.008429997387057181, "ibn_sina": 0.00656528856633003, "einstein": 0.9743752165490657, "socrates": 0.003514145736653205}, "confidence": 0.911914756019589, "choice": "einstein"}}, "usage": {"input_tokens": 387, "output_tokens": 1}, "latency_ms": 210.80553900173982, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 387, "compiled_prompt_sha256": "b429c3380c2f0253b5f360120a373ebdfca865dff32f4f83aa8149409771f11a", "compiled_ids_sha256": "8d69cf61a4b78ca20ed2eeb182851cbb6fd5610240665d51b41f2bdd84be84ae", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.125, 14.1875, 15.625, 15.375, 20.375, 14.75], "runtime_probabilities": [0.005113052204251289, 0.0020022999960929155, 0.008429997600615025, 0.00656528864055872, 0.9743752479553223, 0.003514145966619253], "decide_seconds": 0.20629963299870724, "probabilities_f64_minus_runtime_max_abs": 3.1406256595367665e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.009844451990970683, "einstein": 0.943322338736569, "darwin": 0.0036215714970773876, "sagan": 0.01957803054194739, "hypatia": 0.017277551311975177, "socrates": 0.006356055921460349}, "confidence": 0.8324759019544503, "choice": "einstein"}}, "usage": {"input_tokens": 387, "output_tokens": 1}, "latency_ms": 210.41340200099512, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 387, "compiled_prompt_sha256": "32569ed4b1860f0ccaa5b47061cf534a8b0cec2159b6ea6dcf44768ba0300d0f", "compiled_ids_sha256": "8db1e8b99c418619af59672c37d9a945d44c227d8d1e9b74ae3dce7467be82c9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 20.5, 14.9375, 16.625, 16.5, 15.5], "runtime_probabilities": [0.009844452142715454, 0.9433223605155945, 0.003621571697294712, 0.019578030332922935, 0.01727755181491375, 0.006356055848300457], "decide_seconds": 0.2057226190008805, "probabilities_f64_minus_runtime_max_abs": 2.1779025471957425e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.9753774123776588, "sagan": 0.006173861454747995, "socrates": 0.003986144550628004, "ibn_sina": 0.008438668079900008, "darwin": 0.004808208135532101, "hypatia": 0.00121570540153303}, "confidence": 0.9152413739178549, "choice": "einstein"}}, "usage": {"input_tokens": 387, "output_tokens": 1}, "latency_ms": 210.69421900028829, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 387, "compiled_prompt_sha256": "d95158421940fd1c60b8d6778bc47021fd09a020413870cd075808a0690e9efb", "compiled_ids_sha256": "d01b6c896e4280e721feef7812f08140bf588566002b821e2329a6254e37509f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.875, 15.8125, 15.375, 16.125, 15.5625, 14.1875], "runtime_probabilities": [0.9753774404525757, 0.006173861678689718, 0.003986144904047251, 0.008438668213784695, 0.00480820843949914, 0.0012157055316492915], "decide_seconds": 0.20622167499823263, "probabilities_f64_minus_runtime_max_abs": 2.807491683576302e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.007426140157679677, "ibn_sina": 0.004504181688955933, "hypatia": 0.0024109147251396316, "einstein": 0.9726324187769506, "sagan": 0.012243635237166888, "darwin": 0.0007827094141071589}, "confidence": 0.9097146356940404, "choice": "einstein"}}, "usage": {"input_tokens": 387, "output_tokens": 1}, "latency_ms": 210.4427019985451, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 387, "compiled_prompt_sha256": "7e24416d18a1e23b747b58650780df0865e19627020d3b2ca042eb06a7df934c", "compiled_ids_sha256": "58234122cb361c455f36bd355cee04afccae04df794d50603124bb2b2e97571b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 15.5, 14.875, 20.875, 16.5, 13.75], "runtime_probabilities": [0.007426139432936907, 0.004504181910306215, 0.0024109147489070892, 0.9726323485374451, 0.012243635021150112, 0.0007827093941159546], "decide_seconds": 0.2059943559979729, "probabilities_f64_minus_runtime_max_abs": 7.023950554518876e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:10Z"} {"id": "c-einstein-01-starter-03", "uid": "c034", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "prompt_sha256_by_order": ["70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c"], "request_sha256": "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "compiled_prompt_sha256_by_order": ["3048ef259ba2c81faa325128892aa67888378d9abaef64bcc7e191dc46c7930d", "7d0d1eb128df1b1634aec02eca91a38321fbf48bd80a3b1ce481bfe00dea0618", "11faceff6b952748ccf699acff210388b55eaa906f0f34d2ae0b2e77649e452a", "ee56731ff21bb019336e14f79c19363ac8b016f5bf686ae6a3bd61984dd9a695", "b645aa52c104637afc161ca122c71a905a2d51cc15286d6d4fea893137dde86f", "6e1c2ab21ae1afd1555a09dc961bb6e14a89b5c4d6efdb6289f36bb9106295c6"], "compiled_tokens_by_order": [343, 343, 343, 343, 343, 343], "choice": "einstein", "probs": {"darwin": 0.0024372904798983124, "einstein": 0.9832731576967055, "hypatia": 0.0016751236161334138, "ibn_sina": 0.005846755916449274, "sagan": 0.006223839275705463, "socrates": 0.0005438330151079539}, "confidence": 0.9398789114542254, "seconds": 0.18599732883315787, "seconds_all_orders": 1.1159839729989471, "seconds_by_order": [0.18615149099787232, 0.185782313998061, 0.18609776100129238, 0.18610348100264673, 0.18570624499989208, 0.18614268099918263], "argmax_stable": true, "spread": {"darwin": 0.018910296296798648, "einstein": 0.034274600370014174, "hypatia": 0.0036850246998389362, "ibn_sina": 0.014603267521904661, "sagan": 0.003329538971963363, "socrates": 0.002974987190869507}, "spread_2dp": {"darwin": 0.02, "einstein": 0.040000000000000036, "hypatia": 0.01, "ibn_sina": 0.02, "sagan": 0.01, "socrates": 0.0}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0024372904798983124, "einstein": 0.9832731576967055, "hypatia": 0.0016751236161334138, "ibn_sina": 0.005846755916449274, "sagan": 0.006223839275705463, "socrates": 0.0005438330151079539}, {"ibn_sina": 0.01772973884693097, "hypatia": 0.004482780003781738, "sagan": 0.0028943003037421, "socrates": 0.002399459228560578, "darwin": 0.004482780003781738, "einstein": 0.968010941613203}, {"hypatia": 0.005120203379840202, "darwin": 0.0037460208160765107, "sagan": 0.0037460208160765107, "ibn_sina": 0.010182740313370102, "einstein": 0.9757380518454126, "socrates": 0.0014669628292239758}, {"ibn_sina": 0.003566428188079513, "einstein": 0.988871691296508, "darwin": 0.0009017342615406785, "sagan": 0.002956672815817282, "hypatia": 0.0033503492274680902, "socrates": 0.000353124210586578}, {"einstein": 0.9545970909264938, "sagan": 0.004152827207613445, "socrates": 0.0025188170259063565, "ibn_sina": 0.017484055601645808, "darwin": 0.019812030558339327, "hypatia": 0.001435178680001266}, {"socrates": 0.003328111401456085, "ibn_sina": 0.0031264713250263097, "hypatia": 0.0035427561455113536, "einstein": 0.9823080899743055, "sagan": 0.006217730671654419, "darwin": 0.0014768404820463105}], "prompt_tokens": 343, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0024372904798983124, "einstein": 0.9832731576967055, "hypatia": 0.0016751236161334138, "ibn_sina": 0.005846755916449274, "sagan": 0.006223839275705463, "socrates": 0.0005438330151079539}, "confidence": 0.9398789114542254, "choice": "einstein"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 186.12341100015328, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 343, "compiled_prompt_sha256": "3048ef259ba2c81faa325128892aa67888378d9abaef64bcc7e191dc46c7930d", "compiled_ids_sha256": "dee894996ae3ae3d47634ad4330ad3215bf4766aff9df1233b15480c06ddd321", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0, 21.0, 14.625, 15.875, 15.9375, 13.5], "runtime_probabilities": [0.0024372905027121305, 0.9832731485366821, 0.001675123698078096, 0.00584675557911396, 0.006223839242011309, 0.0005438330117613077], "decide_seconds": 0.18192797400115523, "probabilities_f64_minus_runtime_max_abs": 9.1600234064515e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.01772973884693097, "hypatia": 0.004482780003781738, "sagan": 0.0028943003037421, "socrates": 0.002399459228560578, "darwin": 0.004482780003781738, "einstein": 0.968010941613203}, "confidence": 0.897954715515452, "choice": "einstein"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 185.75861399949645, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 343, "compiled_prompt_sha256": "7d0d1eb128df1b1634aec02eca91a38321fbf48bd80a3b1ce481bfe00dea0618", "compiled_ids_sha256": "9c06a7708d2042ac30c7d558072e3af4c5b5e55520e466b3ffd8ae4667ccbd41", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 15.125, 14.6875, 14.5, 15.125, 20.5], "runtime_probabilities": [0.0177297405898571, 0.004482780583202839, 0.002894300501793623, 0.002399459481239319, 0.004482780583202839, 0.9680109620094299], "decide_seconds": 0.18169672499789158, "probabilities_f64_minus_runtime_max_abs": 2.0396226929619843e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.005120203379840202, "darwin": 0.0037460208160765107, "sagan": 0.0037460208160765107, "ibn_sina": 0.010182740313370102, "einstein": 0.9757380518454126, "socrates": 0.0014669628292239758}, "confidence": 0.9167797458039586, "choice": "einstein"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 186.07552100002067, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 343, "compiled_prompt_sha256": "11faceff6b952748ccf699acff210388b55eaa906f0f34d2ae0b2e77649e452a", "compiled_ids_sha256": "29aa305dd02ddde621698702f0ad4b9afa944197055976ac3b0a3b010b67016d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.25, 14.9375, 14.9375, 15.9375, 20.5, 14.0], "runtime_probabilities": [0.005120203830301762, 0.003746021306142211, 0.003746021306142211, 0.010182742029428482, 0.9757381081581116, 0.0014669628581032157], "decide_seconds": 0.18206620199998724, "probabilities_f64_minus_runtime_max_abs": 5.631269894834645e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.003566428188079513, "einstein": 0.988871691296508, "darwin": 0.0009017342615406785, "sagan": 0.002956672815817282, "hypatia": 0.0033503492274680902, "socrates": 0.000353124210586578}, "confidence": 0.9572444367270267, "choice": "einstein"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 186.08021199906943, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 343, "compiled_prompt_sha256": "ee56731ff21bb019336e14f79c19363ac8b016f5bf686ae6a3bd61984dd9a695", "compiled_ids_sha256": "b4bc04d940bf7bd555bc38846260b794f9414ea8300db287ad490da47b828946", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5, 21.125, 14.125, 15.3125, 15.4375, 13.1875], "runtime_probabilities": [0.0035664280876517296, 0.9888716340065002, 0.0009017342235893011, 0.002956672804430127, 0.0033503491431474686, 0.00035312416730448604], "decide_seconds": 0.18190168400178663, "probabilities_f64_minus_runtime_max_abs": 5.729000773779802e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.9545970909264938, "sagan": 0.004152827207613445, "socrates": 0.0025188170259063565, "ibn_sina": 0.017484055601645808, "darwin": 0.019812030558339327, "hypatia": 0.001435178680001266}, "confidence": 0.8660317401766258, "choice": "einstein"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 185.6829749995086, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 343, "compiled_prompt_sha256": "b645aa52c104637afc161ca122c71a905a2d51cc15286d6d4fea893137dde86f", "compiled_ids_sha256": "6cc6d697b909a6695bb779df4e748c793556a899be9c8812c76e593c88709b55", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.75, 15.3125, 14.8125, 16.75, 16.875, 14.25], "runtime_probabilities": [0.9545971155166626, 0.004152827430516481, 0.0025188170839101076, 0.01748405583202839, 0.019812030717730522, 0.0014351786812767386], "decide_seconds": 0.1816627160005737, "probabilities_f64_minus_runtime_max_abs": 2.459016879008402e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.003328111401456085, "ibn_sina": 0.0031264713250263097, "hypatia": 0.0035427561455113536, "einstein": 0.9823080899743055, "sagan": 0.006217730671654419, "darwin": 0.0014768404820463105}, "confidence": 0.9353926370877181, "choice": "einstein"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 186.11937099922216, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 343, "compiled_prompt_sha256": "6e1c2ab21ae1afd1555a09dc961bb6e14a89b5c4d6efdb6289f36bb9106295c6", "compiled_ids_sha256": "e49b95bdfc653d05318070b9204906b2f7090777af85b775a6215ee2a0fcc56d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.1875, 15.125, 15.25, 20.875, 15.8125, 14.375], "runtime_probabilities": [0.003328111255541444, 0.0031264713034033775, 0.003542755963280797, 0.982308030128479, 0.0062177302315831184, 0.0014768403489142656], "decide_seconds": 0.18211295199944288, "probabilities_f64_minus_runtime_max_abs": 5.984582651663573e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:11Z"} {"id": "c-einstein-01-starter-03", "uid": "c035", "task": "kperm", "label": "einstein", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "prompt_sha256_by_order": ["36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156"], "request_sha256": "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "compiled_prompt_sha256_by_order": ["8f4e67e82afe236dfe30a2d5eeec24f6f2778b73186fa1f61532a001b7adee48", "56ba92fcf591d9e229f3f8ad5237c5869776100cf3a70c6bec64ce9064009b29", "c78dcb949364652ccb81e7ece0a7c58511b69ae2b059a4200cc368f7cd710b86", "8b89de27ed879d98c030e2f398520cd0febf51c68c5518567109e07953145424", "c0960b215a98d11febcac1885a2697564706bea112c92b9221476c4435d5d8b9", "af19bfbc7bf8d777b190a6366cb8b8ef37d3d9885df8ab53705093e1856fea10"], "compiled_tokens_by_order": [278, 278, 278, 278, 278, 278], "choice": "einstein", "probs": {"darwin": 0.004724820579627463, "einstein": 0.7012255484153789, "hypatia": 0.25796646288618774, "ibn_sina": 0.009396427723758658, "sagan": 0.02399461877884033, "socrates": 0.0026921216162069327}, "confidence": 0.5685832111897112, "seconds": 0.1728329328319281, "seconds_all_orders": 1.0369975969915686, "seconds_by_order": [0.17294770299849915, 0.17278069500025595, 0.17285972499666968, 0.17297903299913742, 0.17269171599764377, 0.1727387249993626], "argmax_stable": false, "spread": {"darwin": 0.015440795072542451, "einstein": 0.655904294992785, "hypatia": 0.6552274020313913, "ibn_sina": 0.01427031434376602, "sagan": 0.027712175573754903, "socrates": 0.006582218641195414}, "spread_2dp": {"darwin": 0.02, "einstein": 0.66, "hypatia": 0.66, "ibn_sina": 0.02, "sagan": 0.03, "socrates": 0.01}, "choices_by_order": ["einstein", "hypatia", "einstein", "einstein", "hypatia", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.004724820579627463, "einstein": 0.7012255484153789, "hypatia": 0.25796646288618774, "ibn_sina": 0.009396427723758658, "sagan": 0.02399461877884033, "socrates": 0.0026921216162069327}, {"ibn_sina": 0.005638806585634924, "hypatia": 0.7385379176731215, "sagan": 0.01053467760835464, "socrates": 0.0025022052544429758, "darwin": 0.0030182356681692393, "einstein": 0.23976815721027667}, {"hypatia": 0.0833105156417302, "darwin": 0.0072795917286127216, "sagan": 0.007749085058292525, "ibn_sina": 0.004147786338016819, "einstein": 0.8956724522030617, "socrates": 0.0018405690302860163}, {"ibn_sina": 0.01653281340927195, "einstein": 0.6203892462322691, "darwin": 0.005042223208143086, "sagan": 0.021228552626861092, "hypatia": 0.33207043417610393, "socrates": 0.0047367303473508865}, {"einstein": 0.43200659348884973, "sagan": 0.03546126063204743, "socrates": 0.00842278767148143, "ibn_sina": 0.01783104164045122, "darwin": 0.016750713440610902, "hypatia": 0.4895276031265593}, {"socrates": 0.003790374477038962, "ibn_sina": 0.003560727296685198, "hypatia": 0.1611708239491188, "einstein": 0.8184929805812043, "sagan": 0.011675175327884283, "darwin": 0.0013099183680684513}], "prompt_tokens": 278, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.004724820579627463, "einstein": 0.7012255484153789, "hypatia": 0.25796646288618774, "ibn_sina": 0.009396427723758658, "sagan": 0.02399461877884033, "socrates": 0.0026921216162069327}, "confidence": 0.5685832111897112, "choice": "einstein"}}, "usage": {"input_tokens": 278, "output_tokens": 1}, "latency_ms": 172.92236399953254, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 278, "compiled_prompt_sha256": "8f4e67e82afe236dfe30a2d5eeec24f6f2778b73186fa1f61532a001b7adee48", "compiled_ids_sha256": "bc3a2e6b1255920114fcfe262b849bf8a7b90aa700f811ca5a13e3dea051bd85", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0, 20.0, 19.0, 15.6875, 16.625, 14.4375], "runtime_probabilities": [0.004724820610135794, 0.7012255787849426, 0.2579664885997772, 0.009396428242325783, 0.02399461902678013, 0.0026921217795461416], "decide_seconds": 0.16944394999882206, "probabilities_f64_minus_runtime_max_abs": 3.0369563752685735e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.005638806585634924, "hypatia": 0.7385379176731215, "sagan": 0.01053467760835464, "socrates": 0.0025022052544429758, "darwin": 0.0030182356681692393, "einstein": 0.23976815721027667}, "confidence": 0.6227645657110741, "choice": "hypatia"}}, "usage": {"input_tokens": 278, "output_tokens": 1}, "latency_ms": 172.758425000211, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 278, "compiled_prompt_sha256": "56ba92fcf591d9e229f3f8ad5237c5869776100cf3a70c6bec64ce9064009b29", "compiled_ids_sha256": "6a4869b0ad66deb5f8da671c706b01b51b4f0ded31173e49dacf6f558a58b9e2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.375, 20.25, 16.0, 14.5625, 14.75, 19.125], "runtime_probabilities": [0.005638806615024805, 0.7385379672050476, 0.010534677654504776, 0.002502205315977335, 0.0030182357877492905, 0.2397681623697281], "decide_seconds": 0.16936596099913004, "probabilities_f64_minus_runtime_max_abs": 4.9531926116230807e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0833105156417302, "darwin": 0.0072795917286127216, "sagan": 0.007749085058292525, "ibn_sina": 0.004147786338016819, "einstein": 0.8956724522030617, "socrates": 0.0018405690302860163}, "confidence": 0.769183842477177, "choice": "einstein"}}, "usage": {"input_tokens": 278, "output_tokens": 1}, "latency_ms": 172.8378249972593, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 278, "compiled_prompt_sha256": "c78dcb949364652ccb81e7ece0a7c58511b69ae2b059a4200cc368f7cd710b86", "compiled_ids_sha256": "d0d27afdbcce55ed017cb26871c2c98e4051b37964751abc8bd378765be94ee5", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 15.3125, 15.375, 14.75, 20.125, 13.9375], "runtime_probabilities": [0.08331050723791122, 0.007279592100530863, 0.00774908484891057, 0.004147786647081375, 0.8956723809242249, 0.001840568846091628], "decide_seconds": 0.16949168000064674, "probabilities_f64_minus_runtime_max_abs": 7.127883683288871e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.01653281340927195, "einstein": 0.6203892462322691, "darwin": 0.005042223208143086, "sagan": 0.021228552626861092, "hypatia": 0.33207043417610393, "socrates": 0.0047367303473508865}, "confidence": 0.517855331037977, "choice": "einstein"}}, "usage": {"input_tokens": 278, "output_tokens": 1}, "latency_ms": 172.95637399729458, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 278, "compiled_prompt_sha256": "8b89de27ed879d98c030e2f398520cd0febf51c68c5518567109e07953145424", "compiled_ids_sha256": "fbf019471853ab3512dfdfd5a3519d1703362c5a14b3eae89f8739cf6298fceb", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.875, 19.5, 14.6875, 16.125, 18.875, 14.625], "runtime_probabilities": [0.016532812267541885, 0.6203892230987549, 0.00504222372546792, 0.021228551864624023, 0.3320704400539398, 0.004736729897558689], "decide_seconds": 0.16942967099748785, "probabilities_f64_minus_runtime_max_abs": 2.3133514215345485e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.43200659348884973, "sagan": 0.03546126063204743, "socrates": 0.00842278767148143, "ibn_sina": 0.01783104164045122, "darwin": 0.016750713440610902, "hypatia": 0.4895276031265593}, "confidence": 0.4356287006864966, "choice": "hypatia"}}, "usage": {"input_tokens": 278, "output_tokens": 1}, "latency_ms": 172.66965599992545, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 278, "compiled_prompt_sha256": "c0960b215a98d11febcac1885a2697564706bea112c92b9221476c4435d5d8b9", "compiled_ids_sha256": "fc7df527b549d3694ff6cc0cf2bb2612d262bb2ac202c3ce0ad19182b48bc9ff", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.125, 16.625, 15.1875, 15.9375, 15.875, 19.25], "runtime_probabilities": [0.4320065975189209, 0.03546126186847687, 0.008422787301242352, 0.017831042408943176, 0.01675071381032467, 0.4895276129245758], "decide_seconds": 0.1692720620012551, "probabilities_f64_minus_runtime_max_abs": 9.798016509243723e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.003790374477038962, "ibn_sina": 0.003560727296685198, "hypatia": 0.1611708239491188, "einstein": 0.8184929805812043, "sagan": 0.011675175327884283, "darwin": 0.0013099183680684513}, "confidence": 0.6874692899217101, "choice": "einstein"}}, "usage": {"input_tokens": 278, "output_tokens": 1}, "latency_ms": 172.71759499999462, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 278, "compiled_prompt_sha256": "af19bfbc7bf8d777b190a6366cb8b8ef37d3d9885df8ab53705093e1856fea10", "compiled_ids_sha256": "5b52528a09013dbddda1f9e47550ee0c9ffc92cfb9f43beb8f1df7a3e7646fb1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.875, 14.8125, 18.625, 20.25, 16.0, 13.8125], "runtime_probabilities": [0.003790374845266342, 0.0035607274621725082, 0.1611708253622055, 0.8184930086135864, 0.011675175279378891, 0.0013099184725433588], "decide_seconds": 0.16933538100056467, "probabilities_f64_minus_runtime_max_abs": 2.8032382082265883e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:12Z"} {"id": "c-hypatia-01-dessert-01", "uid": "c036", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "prompt_sha256_by_order": ["53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816"], "request_sha256": "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "compiled_prompt_sha256_by_order": ["4231600771fa3f6d01b11d9780088dab4c81fa8caa77568d942fc82a7a6b4f8c", "3e2f4297679effe8564f1df866abfb4406e82fc1c02552e3e8e74020d5222e3c", "0955081e6fd04107249a73b24edf436a83412f5c8545df9a1099bc8971404ccd", "845af03cd605ee29c93cdb5eb1a4d16ea4d5d00a134d31d4927d0ca50a95c13a", "820c72d1ae90bd487c7673c58e8ba149931905682bacacf34aef5a7b3f1e5485", "d7cbff19edacb7ae424f7fa26d318d9443bc71d83980e0de0903b52472c5bb5c"], "compiled_tokens_by_order": [395, 395, 395, 395, 395, 395], "choice": "ibn_sina", "probs": {"darwin": 0.02714854666916764, "einstein": 0.023958508345213778, "hypatia": 0.20060213434291654, "ibn_sina": 0.5452931365344502, "sagan": 0.065125977871779, "socrates": 0.13787169623647283}, "confidence": 0.27930272777967313, "seconds": 0.2116223456669104, "seconds_all_orders": 1.2697340740014624, "seconds_by_order": [0.211718042002758, 0.21141487400018377, 0.21172904099876177, 0.21126687500145636, 0.2117684309996548, 0.2118368109986477], "argmax_stable": false, "spread": {"darwin": 0.055202368309207606, "einstein": 0.06489569575819099, "hypatia": 0.2554517020822683, "ibn_sina": 0.4871665779412149, "sagan": 0.052280443379536856, "socrates": 0.32382446457033787}, "spread_2dp": {"darwin": 0.06, "einstein": 0.06, "hypatia": 0.26, "ibn_sina": 0.48999999999999994, "sagan": 0.05, "socrates": 0.31999999999999995}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.02714854666916764, "einstein": 0.023958508345213778, "hypatia": 0.20060213434291654, "ibn_sina": 0.5452931365344502, "sagan": 0.065125977871779, "socrates": 0.13787169623647283}, {"ibn_sina": 0.4060481957782442, "hypatia": 0.09060159895198061, "sagan": 0.07056059621132389, "socrates": 0.35833627507436394, "darwin": 0.048495541281141794, "einstein": 0.02595779270294562}, {"hypatia": 0.13748051596887667, "darwin": 0.011285087964118139, "sagan": 0.04463338872299481, "ibn_sina": 0.6161449258021524, "einstein": 0.08338614804823602, "socrates": 0.10706993349362195}, {"ibn_sina": 0.6326256774001356, "einstein": 0.024529559495196782, "darwin": 0.0588433542449236, "sagan": 0.045827250364494566, "hypatia": 0.14115786871230115, "socrates": 0.09701628978294821}, {"einstein": 0.038699083848575386, "sagan": 0.018280152831787036, "socrates": 0.1350730747711004, "ibn_sina": 0.6859576743280048, "darwin": 0.04969060726410086, "hypatia": 0.07229940695643158}, {"socrates": 0.42084075435328605, "ibn_sina": 0.1987910963867899, "hypatia": 0.3277511090386999, "einstein": 0.018490452290045038, "sagan": 0.03048560199546314, "darwin": 0.003640985935715994}], "prompt_tokens": 395, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.02714854666916764, "einstein": 0.023958508345213778, "hypatia": 0.20060213434291654, "ibn_sina": 0.5452931365344502, "sagan": 0.065125977871779, "socrates": 0.13787169623647283}, "confidence": 0.27930272777967313, "choice": "ibn_sina"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 211.6898320018663, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "4231600771fa3f6d01b11d9780088dab4c81fa8caa77568d942fc82a7a6b4f8c", "compiled_ids_sha256": "7ef9a0540d1930f31f588689d49c705e105a918541d1c90201320219a03b21c8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 16.25, 18.375, 19.375, 17.25, 18.0], "runtime_probabilities": [0.027148546651005745, 0.023958507925271988, 0.20060215890407562, 0.5452931523323059, 0.0651259794831276, 0.13787169754505157], "decide_seconds": 0.2070995279973431, "probabilities_f64_minus_runtime_max_abs": 2.45611590787842e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.4060481957782442, "hypatia": 0.09060159895198061, "sagan": 0.07056059621132389, "socrates": 0.35833627507436394, "darwin": 0.048495541281141794, "einstein": 0.02595779270294562}, "confidence": 0.22986546316708445, "choice": "ibn_sina"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 211.3907040002232, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "3e2f4297679effe8564f1df866abfb4406e82fc1c02552e3e8e74020d5222e3c", "compiled_ids_sha256": "1baafb93faf4b18a2b85ac937d9b7d0684e5f91e6451614c698e0f74463d5d5a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.0, 17.5, 17.25, 18.875, 16.875, 16.25], "runtime_probabilities": [0.4060481786727905, 0.0906016007065773, 0.07056058943271637, 0.35833626985549927, 0.048495542258024216, 0.02595779113471508], "decide_seconds": 0.20693893800125807, "probabilities_f64_minus_runtime_max_abs": 1.7105453664623127e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.13748051596887667, "darwin": 0.011285087964118139, "sagan": 0.04463338872299481, "ibn_sina": 0.6161449258021524, "einstein": 0.08338614804823602, "socrates": 0.10706993349362195}, "confidence": 0.32639310933923815, "choice": "ibn_sina"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 211.7045019986108, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "0955081e6fd04107249a73b24edf436a83412f5c8545df9a1099bc8971404ccd", "compiled_ids_sha256": "2408a912179e8d8842eda565b9269ab95eb249ac1a73d511b205466a87a119fc", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 15.625, 17.0, 19.625, 17.625, 17.875], "runtime_probabilities": [0.13748052716255188, 0.011285088025033474, 0.04463339224457741, 0.6161449551582336, 0.0833861455321312, 0.10706992447376251], "decide_seconds": 0.20703120799953467, "probabilities_f64_minus_runtime_max_abs": 2.935608123255662e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.6326256774001356, "einstein": 0.024529559495196782, "darwin": 0.0588433542449236, "sagan": 0.045827250364494566, "hypatia": 0.14115786871230115, "socrates": 0.09701628978294821}, "confidence": 0.335128786848643, "choice": "ibn_sina"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 211.24280500225723, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "845af03cd605ee29c93cdb5eb1a4d16ea4d5d00a134d31d4927d0ca50a95c13a", "compiled_ids_sha256": "e435ee4c0012e1a2937895a9c9c205c811c1f490c1f4f0d18631555c828df116", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.5, 16.25, 17.125, 16.875, 18.0, 17.625], "runtime_probabilities": [0.6326256394386292, 0.024529557675123215, 0.058843351900577545, 0.04582725092768669, 0.141157865524292, 0.09701628983020782], "decide_seconds": 0.20675752099850797, "probabilities_f64_minus_runtime_max_abs": 3.796150649382213e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.038699083848575386, "sagan": 0.018280152831787036, "socrates": 0.1350730747711004, "ibn_sina": 0.6859576743280048, "darwin": 0.04969060726410086, "hypatia": 0.07229940695643158}, "confidence": 0.40445687771198213, "choice": "ibn_sina"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 211.74373099711374, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "820c72d1ae90bd487c7673c58e8ba149931905682bacacf34aef5a7b3f1e5485", "compiled_ids_sha256": "1add6aea2da9fb9d0a9c8833a85e5962baa4522d3edbf1acdc58ee030bbf50e2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 16.0, 18.0, 19.625, 17.0, 17.375], "runtime_probabilities": [0.038699086755514145, 0.01828015223145485, 0.13507306575775146, 0.685957670211792, 0.0496906079351902, 0.07229940593242645], "decide_seconds": 0.20723060700038332, "probabilities_f64_minus_runtime_max_abs": 9.013348928332832e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.42084075435328605, "ibn_sina": 0.1987910963867899, "hypatia": 0.3277511090386999, "einstein": 0.018490452290045038, "sagan": 0.03048560199546314, "darwin": 0.003640985935715994}, "confidence": 0.30144980649800335, "choice": "socrates"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 211.81156099919463, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "d7cbff19edacb7ae424f7fa26d318d9443bc71d83980e0de0903b52472c5bb5c", "compiled_ids_sha256": "7eafaa910c92229237d1aa531c2512571a888d4741a40cfb06fa2e69ebbf824d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.5, 18.75, 19.25, 16.375, 16.875, 14.75], "runtime_probabilities": [0.42084071040153503, 0.19879108667373657, 0.3277510702610016, 0.018490450456738472, 0.030485600233078003, 0.003640985582023859], "decide_seconds": 0.20705639800144127, "probabilities_f64_minus_runtime_max_abs": 4.395175101645776e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:14Z"} {"id": "c-hypatia-01-dessert-02", "uid": "c037", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "prompt_sha256_by_order": ["60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b"], "request_sha256": "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "compiled_prompt_sha256_by_order": ["64cf4ddb66e4162fd4413e7c5eecfae78ed5607c7644b898fafd96f028bf8081", "126fd02f52a3255d24453db4c6d6615c2078c7e398d4dc621c4794f1dbc27888", "d4b5d824760b9b2b66b91bdb06cb7be6f8c7c4739faf8eeb587a17f85422db5d", "41469d816178843dac72662e56417911a5b5854d202c276b41a222ce9977785f", "48c60228de038271c453cf978f22cde2a0fb1bcf131fe7c1e2364daa12a7274c", "1fb528d3c58e26b2dbf098c2beb55874c4c20fc362e400b8bc58ba0599f2670d"], "compiled_tokens_by_order": [323, 323, 323, 323, 323, 323], "choice": "einstein", "probs": {"darwin": 0.08733966420397367, "einstein": 0.44354741339437553, "hypatia": 0.03213046685945926, "ibn_sina": 0.028355037482069786, "sagan": 0.3914292184699455, "socrates": 0.017198199590176235}, "confidence": 0.3179781212219174, "seconds": 0.18387664516558289, "seconds_all_orders": 1.1032598709934973, "seconds_by_order": [0.18384786899696337, 0.1837363499980711, 0.18402263700045296, 0.1841943559993524, 0.1838312389991188, 0.1836274199995387], "argmax_stable": false, "spread": {"darwin": 0.05606676344289216, "einstein": 0.2892497839829794, "hypatia": 0.08104978164834388, "ibn_sina": 0.06061814398537352, "sagan": 0.3000847808421818, "socrates": 0.140601082378695}, "spread_2dp": {"darwin": 0.06, "einstein": 0.29000000000000004, "hypatia": 0.08, "ibn_sina": 0.06, "sagan": 0.29999999999999993, "socrates": 0.13999999999999999}, "choices_by_order": ["einstein", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.08733966420397367, "einstein": 0.44354741339437553, "hypatia": 0.03213046685945926, "ibn_sina": 0.028355037482069786, "sagan": 0.3914292184699455, "socrates": 0.017198199590176235}, {"ibn_sina": 0.07476544660823105, "hypatia": 0.03116687507527288, "sagan": 0.5524460792498235, "socrates": 0.04534753565500005, "darwin": 0.06598027505211784, "einstein": 0.23029378835955472}, {"hypatia": 0.09632731583882478, "darwin": 0.03127290076108151, "sagan": 0.5543254294985502, "ibn_sina": 0.05156029668128931, "einstein": 0.23107721809992227, "socrates": 0.03543683912033188}, {"ibn_sina": 0.0889731814674433, "einstein": 0.2740566926578604, "darwin": 0.042027955016176934, "sagan": 0.5120053081652927, "hypatia": 0.06929238339919906, "socrates": 0.01364447929402757}, {"einstein": 0.15429762941139616, "sagan": 0.6915139993121273, "socrates": 0.06432082142804063, "ibn_sina": 0.030383004687433945, "darwin": 0.04420701097052095, "hypatia": 0.015277534190480904}, {"socrates": 0.15424556167272258, "ibn_sina": 0.044192093313629305, "hypatia": 0.09355466227910211, "einstein": 0.25430793844090616, "sagan": 0.4192829074154208, "darwin": 0.0344168368782191}], "prompt_tokens": 323, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.08733966420397367, "einstein": 0.44354741339437553, "hypatia": 0.03213046685945926, "ibn_sina": 0.028355037482069786, "sagan": 0.3914292184699455, "socrates": 0.017198199590176235}, "confidence": 0.3179781212219174, "choice": "einstein"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 183.82110899983672, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "64cf4ddb66e4162fd4413e7c5eecfae78ed5607c7644b898fafd96f028bf8081", "compiled_ids_sha256": "2edffcb25f65010db4bf8ec46a80e7f11423c9168ec99e3b3ac49296338f2c48", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 19.25, 16.625, 16.5, 19.125, 16.0], "runtime_probabilities": [0.08733965456485748, 0.44354739785194397, 0.032130468636751175, 0.028355035930871964, 0.39142921566963196, 0.017198199406266212], "decide_seconds": 0.17987204900055076, "probabilities_f64_minus_runtime_max_abs": 1.554243156398627e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.07476544660823105, "hypatia": 0.03116687507527288, "sagan": 0.5524460792498235, "socrates": 0.04534753565500005, "darwin": 0.06598027505211784, "einstein": 0.23029378835955472}, "confidence": 0.2813661018232041, "choice": "sagan"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 183.71385000136797, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "126fd02f52a3255d24453db4c6d6615c2078c7e398d4dc621c4794f1dbc27888", "compiled_ids_sha256": "a9dcef5a25e17758145e02113f8a18e8cf23c6d68f0a239da3291718a3d2435b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 16.5, 19.375, 16.875, 17.25, 18.5], "runtime_probabilities": [0.07476544380187988, 0.03116687573492527, 0.5524460673332214, 0.045347534120082855, 0.06598027795553207, 0.23029378056526184], "decide_seconds": 0.17989059899991844, "probabilities_f64_minus_runtime_max_abs": 1.1916602038475332e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.09632731583882478, "darwin": 0.03127290076108151, "sagan": 0.5543254294985502, "ibn_sina": 0.05156029668128931, "einstein": 0.23107721809992227, "socrates": 0.03543683912033188}, "confidence": 0.2908728324903782, "choice": "sagan"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 183.99924700133852, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "d4b5d824760b9b2b66b91bdb06cb7be6f8c7c4739faf8eeb587a17f85422db5d", "compiled_ids_sha256": "55e8c17d426b8b4a5a6226037d4a13fa22099e749ae06aab91c690a6d167adaa", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 16.625, 19.5, 17.125, 18.625, 16.75], "runtime_probabilities": [0.09632730484008789, 0.031272899359464645, 0.5543254017829895, 0.05156029388308525, 0.23107720911502838, 0.03543683513998985], "decide_seconds": 0.1800604790005309, "probabilities_f64_minus_runtime_max_abs": 2.771556073533077e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0889731814674433, "einstein": 0.2740566926578604, "darwin": 0.042027955016176934, "sagan": 0.5120053081652927, "hypatia": 0.06929238339919906, "socrates": 0.01364447929402757}, "confidence": 0.28030257359238075, "choice": "sagan"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 184.17181599943433, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "41469d816178843dac72662e56417911a5b5854d202c276b41a222ce9977785f", "compiled_ids_sha256": "ad858089a65d8558ae08ba6a32b72aa795e8595e040ade53729b92e1dfa8ee88", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 18.75, 16.875, 19.375, 17.375, 15.75], "runtime_probabilities": [0.08897317200899124, 0.27405670285224915, 0.04202795401215553, 0.5120052695274353, 0.06929238140583038, 0.013644478283822536], "decide_seconds": 0.18031356600113213, "probabilities_f64_minus_runtime_max_abs": 3.863785735802594e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.15429762941139616, "sagan": 0.6915139993121273, "socrates": 0.06432082142804063, "ibn_sina": 0.030383004687433945, "darwin": 0.04420701097052095, "hypatia": 0.015277534190480904}, "confidence": 0.4263504310628703, "choice": "sagan"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 183.8082590002159, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "48c60228de038271c453cf978f22cde2a0fb1bcf131fe7c1e2364daa12a7274c", "compiled_ids_sha256": "aad3d2601ba3acda91bd63406dc2e8d9f7609b44ed309b8847048d312197eeaa", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 19.625, 17.25, 16.5, 16.875, 15.8125], "runtime_probabilities": [0.1542976349592209, 0.6915140151977539, 0.06432082504034042, 0.030383005738258362, 0.04420701041817665, 0.01527753472328186], "decide_seconds": 0.18002233900188003, "probabilities_f64_minus_runtime_max_abs": 1.5885626591938262e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.15424556167272258, "ibn_sina": 0.044192093313629305, "hypatia": 0.09355466227910211, "einstein": 0.25430793844090616, "sagan": 0.4192829074154208, "darwin": 0.0344168368782191}, "confidence": 0.17599697837436434, "choice": "sagan"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 183.6055200001283, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "1fb528d3c58e26b2dbf098c2beb55874c4c20fc362e400b8bc58ba0599f2670d", "compiled_ids_sha256": "2e7101de9cfe6c101b9944ff7428dc9155907f17e6fd68c36d0d043c55b402cc", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 16.875, 17.625, 18.625, 19.125, 16.625], "runtime_probabilities": [0.15424557030200958, 0.04419209063053131, 0.09355466067790985, 0.25430792570114136, 0.4192828834056854, 0.03441683575510979], "decide_seconds": 0.17981861000225763, "probabilities_f64_minus_runtime_max_abs": 2.4009735366536944e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:15Z"} {"id": "c-hypatia-01-dessert-02", "uid": "c038", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "prompt_sha256_by_order": ["cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70"], "request_sha256": "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "compiled_prompt_sha256_by_order": ["b813667369f4e70abaa4bbd36b3b8b690e78c38b9ec85f4f3d5388a491e9ce79", "2f940aa121ab8e63c8413de9a7abfaaf942772de826f2b4bbae41625a4667b75", "4b7a7ef91594b7dff62c49741edbfe626704376c8c9f2323971eb2a90b1660eb", "69047759db4cdc0a6aac6d5730c13eb5159344815b3f54af7c431b13ededc1d5", "fee024091133f61a5a8e7910bb908002435ca23f94302179c6ce66d24dc46941", "ee0f15ee9be62c89613213d7c0e9b66d4048e351ebdb9dfc8b68082c09a24ed8"], "compiled_tokens_by_order": [274, 274, 274, 274, 274, 274], "choice": "sagan", "probs": {"darwin": 0.049591490291443316, "einstein": 0.008617708829528567, "hypatia": 0.049591490291443316, "ibn_sina": 0.007605101349434881, "sagan": 0.8790301982311801, "socrates": 0.005564011006969755}, "confidence": 0.7107669610678722, "seconds": 0.17235657300019133, "seconds_all_orders": 1.034139438001148, "seconds_by_order": [0.17261597600008827, 0.17250042699743062, 0.17205423000268638, 0.1726627060015744, 0.17185436200088589, 0.17245173699848237], "argmax_stable": true, "spread": {"darwin": 0.04458074949804192, "einstein": 0.025108939758724642, "hypatia": 0.046561706914487966, "ibn_sina": 0.013025177833091685, "sagan": 0.11196988011973141, "socrates": 0.026162415270592293}, "spread_2dp": {"darwin": 0.04, "einstein": 0.03, "hypatia": 0.05, "ibn_sina": 0.02, "sagan": 0.12, "socrates": 0.03}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.049591490291443316, "einstein": 0.008617708829528567, "hypatia": 0.049591490291443316, "ibn_sina": 0.007605101349434881, "sagan": 0.8790301982311801, "socrates": 0.005564011006969755}, {"ibn_sina": 0.00993427723723554, "hypatia": 0.003029783376955348, "sagan": 0.9519296397282212, "socrates": 0.007736822891607472, "darwin": 0.017435199528744925, "einstein": 0.00993427723723554}, {"hypatia": 0.012033334448756405, "darwin": 0.0093715702916516, "sagan": 0.9559261578879078, "ibn_sina": 0.006440979786888129, "einstein": 0.0093715702916516, "socrates": 0.0068563872931443485}, {"ibn_sina": 0.0168944261112548, "einstein": 0.0037696560038200985, "darwin": 0.03576550035818803, "sagan": 0.9224044115461594, "hypatia": 0.0168944261112548, "socrates": 0.004271579869322819}, {"einstein": 0.028878595762544743, "sagan": 0.8439562777681764, "socrates": 0.0224906729938724, "ibn_sina": 0.013641282728354536, "darwin": 0.053952319789693524, "hypatia": 0.037080850957358365}, {"socrates": 0.030433995139915113, "ibn_sina": 0.0038692482781631167, "hypatia": 0.03907802329100221, "einstein": 0.013505003478357992, "sagan": 0.8894117105656043, "darwin": 0.02370201924695722}], "prompt_tokens": 274, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.049591490291443316, "einstein": 0.008617708829528567, "hypatia": 0.049591490291443316, "ibn_sina": 0.007605101349434881, "sagan": 0.8790301982311801, "socrates": 0.005564011006969755}, "confidence": 0.7107669610678722, "choice": "sagan"}}, "usage": {"input_tokens": 274, "output_tokens": 1}, "latency_ms": 172.59210600241204, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 274, "compiled_prompt_sha256": "b813667369f4e70abaa4bbd36b3b8b690e78c38b9ec85f4f3d5388a491e9ce79", "compiled_ids_sha256": "1356d08876a1c4cf2b6410ee614660ec7bf937bb468b506992fcb7bf08d22f64", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 15.375, 17.125, 15.25, 20.0, 14.9375], "runtime_probabilities": [0.04959149286150932, 0.008617709390819073, 0.04959149286150932, 0.0076051014475524426, 0.879030168056488, 0.005564010702073574], "decide_seconds": 0.16917914200166706, "probabilities_f64_minus_runtime_max_abs": 3.017469207833301e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.00993427723723554, "hypatia": 0.003029783376955348, "sagan": 0.9519296397282212, "socrates": 0.007736822891607472, "darwin": 0.017435199528744925, "einstein": 0.00993427723723554}, "confidence": 0.8524858097317161, "choice": "sagan"}}, "usage": {"input_tokens": 274, "output_tokens": 1}, "latency_ms": 172.4793569992471, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 274, "compiled_prompt_sha256": "2f940aa121ab8e63c8413de9a7abfaaf942772de826f2b4bbae41625a4667b75", "compiled_ids_sha256": "e3628eef31f404db4947d62516ea71b5897d6d12c87f09aa96e9df8effe4ea81", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 14.75, 20.5, 15.6875, 16.5, 15.9375], "runtime_probabilities": [0.009934277273714542, 0.003029783256351948, 0.9519295692443848, 0.007736823055893183, 0.01743519864976406, 0.009934277273714542], "decide_seconds": 0.16915754300134722, "probabilities_f64_minus_runtime_max_abs": 7.048383643404321e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.012033334448756405, "darwin": 0.0093715702916516, "sagan": 0.9559261578879078, "ibn_sina": 0.006440979786888129, "einstein": 0.0093715702916516, "socrates": 0.0068563872931443485}, "confidence": 0.8602123842379839, "choice": "sagan"}}, "usage": {"input_tokens": 274, "output_tokens": 1}, "latency_ms": 172.03203100143583, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 274, "compiled_prompt_sha256": "4b7a7ef91594b7dff62c49741edbfe626704376c8c9f2323971eb2a90b1660eb", "compiled_ids_sha256": "ea6455a950ef6d665f631c4a2ef5673566a510a7f3cc337c4d782173ebd523c2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 16.0, 20.625, 15.625, 16.0, 15.6875], "runtime_probabilities": [0.01203333493322134, 0.009371571242809296, 0.9559261798858643, 0.006440979894250631, 0.009371571242809296, 0.006856387015432119], "decide_seconds": 0.16859198800011654, "probabilities_f64_minus_runtime_max_abs": 2.1997956456409895e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0168944261112548, "einstein": 0.0037696560038200985, "darwin": 0.03576550035818803, "sagan": 0.9224044115461594, "hypatia": 0.0168944261112548, "socrates": 0.004271579869322819}, "confidence": 0.7902298446391804, "choice": "sagan"}}, "usage": {"input_tokens": 274, "output_tokens": 1}, "latency_ms": 172.64139599865302, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 274, "compiled_prompt_sha256": "69047759db4cdc0a6aac6d5730c13eb5159344815b3f54af7c431b13ededc1d5", "compiled_ids_sha256": "bc26a3e9a780ff4dad9df25142277a157aedbffcc82d966cf5e573f8cc815d17", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 14.75, 17.0, 20.25, 16.25, 14.875], "runtime_probabilities": [0.016894428059458733, 0.003769656177610159, 0.035765502601861954, 0.9224044680595398, 0.016894428059458733, 0.004271580372005701], "decide_seconds": 0.16927701199892908, "probabilities_f64_minus_runtime_max_abs": 5.651338041978704e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.028878595762544743, "sagan": 0.8439562777681764, "socrates": 0.0224906729938724, "ibn_sina": 0.013641282728354536, "darwin": 0.053952319789693524, "hypatia": 0.037080850957358365}, "confidence": 0.6265316963399234, "choice": "sagan"}}, "usage": {"input_tokens": 274, "output_tokens": 1}, "latency_ms": 171.8331819974992, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 274, "compiled_prompt_sha256": "fee024091133f61a5a8e7910bb908002435ca23f94302179c6ce66d24dc46941", "compiled_ids_sha256": "98c41af6d43a5b89eef460919ec169e6f17c46828190e5d394cace8da14cbb84", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 20.125, 16.5, 16.0, 17.375, 17.0], "runtime_probabilities": [0.028878595679998398, 0.8439562916755676, 0.022490672767162323, 0.01364128291606903, 0.05395231768488884, 0.03708085045218468], "decide_seconds": 0.16852563799693598, "probabilities_f64_minus_runtime_max_abs": 1.3907391238632272e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.030433995139915113, "ibn_sina": 0.0038692482781631167, "hypatia": 0.03907802329100221, "einstein": 0.013505003478357992, "sagan": 0.8894117105656043, "darwin": 0.02370201924695722}, "confidence": 0.7178529492440242, "choice": "sagan"}}, "usage": {"input_tokens": 274, "output_tokens": 1}, "latency_ms": 172.43121700084885, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 274, "compiled_prompt_sha256": "ee0f15ee9be62c89613213d7c0e9b66d4048e351ebdb9dfc8b68082c09a24ed8", "compiled_ids_sha256": "9e6f6d29157bceafa71e5e20a62b05c99a10d78c1f78f4112053f0f3eee46ce7", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 14.6875, 17.0, 15.9375, 20.125, 16.5], "runtime_probabilities": [0.03043399564921856, 0.003869248554110527, 0.03907802700996399, 0.013505004346370697, 0.8894117474555969, 0.023702019825577736], "decide_seconds": 0.16912076299922774, "probabilities_f64_minus_runtime_max_abs": 3.6889992616195855e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:16Z"} {"id": "c-hypatia-01-dessert-03", "uid": "c039", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "prompt_sha256_by_order": ["a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e"], "request_sha256": "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "compiled_prompt_sha256_by_order": ["21829cbfc35a4be0d7b256334b6e53727cf63ff22411b3b96a97b55be3b983e1", "d14223f8e924618fd2fa5162ab6bfe1f68c2a20982377dc8992042d6356ac0bc", "90524ea2154ee3db8ecd561b5c8cd5cb035d53cb94f0bb90fe604bbbd1b75f2f", "a252ebbab08bae3d0a4a48241c07c96c8f8536db39c9c0d31d670d252e85d551", "7601ee6d28c81e741632eab0d2d4f312a74d4ad024c7beb8c48c484ba9059fc0", "0a8c432afe9dacff969a9d99c9a7d6e411357522787a2c272bee56910b8c2d8e"], "compiled_tokens_by_order": [298, 298, 298, 298, 298, 298], "choice": "ibn_sina", "probs": {"darwin": 0.0762438232819455, "einstein": 0.024752745349317592, "hypatia": 0.1614081751544924, "ibn_sina": 0.56336988742725, "sagan": 0.14244221462567142, "socrates": 0.031783154161323074}, "confidence": 0.2785562442045576, "seconds": 0.1825102424997264, "seconds_all_orders": 1.0950614549983584, "seconds_by_order": [0.21835383000143338, 0.17487177799921483, 0.17473203999907128, 0.17463564099671203, 0.1765107960018213, 0.17595737000010558], "argmax_stable": false, "spread": {"darwin": 0.2989652442758684, "einstein": 0.18510119767755234, "hypatia": 0.2617073363045431, "ibn_sina": 0.49524028703869816, "sagan": 0.24722326585059293, "socrates": 0.06833821348437366}, "spread_2dp": {"darwin": 0.29000000000000004, "einstein": 0.19, "hypatia": 0.26, "ibn_sina": 0.49000000000000005, "sagan": 0.25, "socrates": 0.06999999999999999}, "choices_by_order": ["ibn_sina", "darwin", "ibn_sina", "ibn_sina", "ibn_sina", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0762438232819455, "einstein": 0.024752745349317592, "hypatia": 0.1614081751544924, "ibn_sina": 0.56336988742725, "sagan": 0.14244221462567142, "socrates": 0.031783154161323074}, {"ibn_sina": 0.23007482540889113, "hypatia": 0.015656788358730155, "sagan": 0.2954219235650108, "socrates": 0.05817202586123165, "darwin": 0.33475689568971817, "einstein": 0.06591754111641805}, {"hypatia": 0.12492513844825967, "darwin": 0.03579165141384976, "sagan": 0.16040705295080043, "ibn_sina": 0.559875627594035, "einstein": 0.09729179564880831, "socrates": 0.021708733944246904}, {"ibn_sina": 0.29563162070745846, "einstein": 0.02749802419462486, "darwin": 0.23023813770763724, "sagan": 0.15824020361525853, "hypatia": 0.26089398958039606, "socrates": 0.02749802419462486}, {"einstein": 0.07946615219295554, "sagan": 0.04819865771441786, "socrates": 0.09004694742862056, "ibn_sina": 0.3142947287984224, "darwin": 0.19062938920231035, "hypatia": 0.27736412466327326}, {"socrates": 0.08748013852767964, "ibn_sina": 0.06812960038855187, "hypatia": 0.2377956709108716, "einstein": 0.20985394302686994, "sagan": 0.2694577966386422, "darwin": 0.1272828505073848}], "prompt_tokens": 298, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0762438232819455, "einstein": 0.024752745349317592, "hypatia": 0.1614081751544924, "ibn_sina": 0.56336988742725, "sagan": 0.14244221462567142, "socrates": 0.031783154161323074}, "confidence": 0.2785562442045576, "choice": "ibn_sina"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 218.3272600013879, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 298, "compiled_prompt_sha256": "21829cbfc35a4be0d7b256334b6e53727cf63ff22411b3b96a97b55be3b983e1", "compiled_ids_sha256": "90ee18b88d0e89b3999ace8ac624b7a7824208a64fcb7a98548c6bbcf7208086", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 16.0, 17.875, 19.125, 17.75, 16.25], "runtime_probabilities": [0.0762438252568245, 0.024752745404839516, 0.1614081859588623, 0.563369870185852, 0.14244221150875092, 0.03178315609693527], "decide_seconds": 0.21452513899930636, "probabilities_f64_minus_runtime_max_abs": 1.7241397975986672e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.23007482540889113, "hypatia": 0.015656788358730155, "sagan": 0.2954219235650108, "socrates": 0.05817202586123165, "darwin": 0.33475689568971817, "einstein": 0.06591754111641805}, "confidence": 0.1771091576086572, "choice": "darwin"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 174.84851799963508, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 298, "compiled_prompt_sha256": "d14223f8e924618fd2fa5162ab6bfe1f68c2a20982377dc8992042d6356ac0bc", "compiled_ids_sha256": "3114c0291676974a6225f6099745086841a066740ff190221b427db42ef5d97f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.0, 15.3125, 18.25, 16.625, 18.375, 16.75], "runtime_probabilities": [0.23007485270500183, 0.015656789764761925, 0.29542192816734314, 0.05817202478647232, 0.33475691080093384, 0.06591754406690598], "decide_seconds": 0.17095788900041953, "probabilities_f64_minus_runtime_max_abs": 2.7296110705066212e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.12492513844825967, "darwin": 0.03579165141384976, "sagan": 0.16040705295080043, "ibn_sina": 0.559875627594035, "einstein": 0.09729179564880831, "socrates": 0.021708733944246904}, "confidence": 0.2704499521814605, "choice": "ibn_sina"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 174.70994999894174, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 298, "compiled_prompt_sha256": "90524ea2154ee3db8ecd561b5c8cd5cb035d53cb94f0bb90fe604bbbd1b75f2f", "compiled_ids_sha256": "0420faee76bd77bf67040c2f86f5e8a070532b9429aff16201fa2a787f279c4c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 16.25, 17.75, 19.0, 17.25, 15.75], "runtime_probabilities": [0.1249251514673233, 0.03579165041446686, 0.16040706634521484, 0.5598756670951843, 0.09729178994894028, 0.021708734333515167], "decide_seconds": 0.1710638980002841, "probabilities_f64_minus_runtime_max_abs": 3.9501149351472975e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.29563162070745846, "einstein": 0.02749802419462486, "darwin": 0.23023813770763724, "sagan": 0.15824020361525853, "hypatia": 0.26089398958039606, "socrates": 0.02749802419462486}, "confidence": 0.14144271254769825, "choice": "ibn_sina"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 174.6117909970053, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 298, "compiled_prompt_sha256": "a252ebbab08bae3d0a4a48241c07c96c8f8536db39c9c0d31d670d252e85d551", "compiled_ids_sha256": "a822ae5c78c7b9e78d08a258d35989a853a092f021918f67a037a062492b64eb", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.375, 16.0, 18.125, 17.75, 18.25, 16.0], "runtime_probabilities": [0.29563164710998535, 0.027498025447130203, 0.2302381545305252, 0.1582402139902115, 0.2608940005302429, 0.027498025447130203], "decide_seconds": 0.170994897998753, "probabilities_f64_minus_runtime_max_abs": 2.6402526887014943e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.07946615219295554, "sagan": 0.04819865771441786, "socrates": 0.09004694742862056, "ibn_sina": 0.3142947287984224, "darwin": 0.19062938920231035, "hypatia": 0.27736412466327326}, "confidence": 0.10724264557631091, "choice": "ibn_sina"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 176.48046599788358, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 298, "compiled_prompt_sha256": "7601ee6d28c81e741632eab0d2d4f312a74d4ad024c7beb8c48c484ba9059fc0", "compiled_ids_sha256": "6b2bf7c2998bfb19b334c181fff999c6c1610dd21800c566bb5f90f23076b154", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 16.625, 17.25, 18.5, 18.0, 18.375], "runtime_probabilities": [0.07946614921092987, 0.04819865897297859, 0.0900469496846199, 0.3142947256565094, 0.19062939286231995, 0.2773641347885132], "decide_seconds": 0.17239106799752335, "probabilities_f64_minus_runtime_max_abs": 1.012523992427461e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.08748013852767964, "ibn_sina": 0.06812960038855187, "hypatia": 0.2377956709108716, "einstein": 0.20985394302686994, "sagan": 0.2694577966386422, "darwin": 0.1272828505073848}, "confidence": 0.06176751545637493, "choice": "sagan"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 175.93408999891835, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 298, "compiled_prompt_sha256": "0a8c432afe9dacff969a9d99c9a7d6e411357522787a2c272bee56910b8c2d8e", "compiled_ids_sha256": "df188a4504ecc2073803b7e34e13219aa8bf8c279cf1ab309a088df460523472", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 17.25, 18.5, 18.375, 18.625, 17.875], "runtime_probabilities": [0.08748014271259308, 0.06812960654497147, 0.23779568076133728, 0.20985394716262817, 0.2694578170776367, 0.12728285789489746], "decide_seconds": 0.17207533100008732, "probabilities_f64_minus_runtime_max_abs": 2.0438994496885243e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:17Z"} {"id": "c-hypatia-01-main-01", "uid": "c040", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "prompt_sha256_by_order": ["6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372"], "request_sha256": "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "compiled_prompt_sha256_by_order": ["a85f301fbcbe5359d5137e8ce9e30a12f45037248d0741358fb7ff25b1b3eafd", "b61c5ac999929626f974063e57c72670796723e26bac9fb12e40f33023fb5628", "081929d17758424a25168221bbd1848ce541fe1930acd1433380e80eb2c816a5", "02542a5ccc36f918849687042233556561fb2d6c471f687f7a179dbb318c5df6", "fbab03d13c25e0f115e70de830408c8fb2989d1e6e51aee8b7ea316b12426d1d", "ceffa6eefc3da573362e5e4396ae6d455b2ab4700f3b030698f74aa75fc83d91"], "compiled_tokens_by_order": [368, 368, 368, 368, 368, 368], "choice": "hypatia", "probs": {"darwin": 0.005584962174309691, "einstein": 0.0033874507920541607, "hypatia": 0.9392462198225568, "ibn_sina": 0.032139238273673335, "sagan": 0.009208045932840025, "socrates": 0.010434083004565957}, "confidence": 0.8278979508740782, "seconds": 0.19402538616668608, "seconds_all_orders": 1.1641523170001165, "seconds_by_order": [0.19391060099951574, 0.19403839899678132, 0.1939377800008515, 0.19392255000275327, 0.19432087699897238, 0.19402211000124225], "argmax_stable": true, "spread": {"darwin": 0.012947605253741534, "einstein": 0.010564070651786923, "hypatia": 0.079096467156894, "ibn_sina": 0.04160217125362636, "sagan": 0.003604265743453838, "socrates": 0.019859666486973668}, "spread_2dp": {"darwin": 0.02, "einstein": 0.01, "hypatia": 0.06999999999999995, "ibn_sina": 0.04, "sagan": 0.0, "socrates": 0.019999999999999997}, "choices_by_order": ["hypatia", "hypatia", "hypatia", "hypatia", "hypatia", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.005584962174309691, "einstein": 0.0033874507920541607, "hypatia": 0.9392462198225568, "ibn_sina": 0.032139238273673335, "sagan": 0.009208045932840025, "socrates": 0.010434083004565957}, {"ibn_sina": 0.011878897974596113, "hypatia": 0.9436577491595973, "sagan": 0.006768392042655814, "socrates": 0.02514762720956024, "darwin": 0.010483090668699494, "einstein": 0.002064242944891072}, {"hypatia": 0.9136606169296246, "darwin": 0.006975884894319403, "sagan": 0.008957213508194192, "ibn_sina": 0.04548848360009997, "einstein": 0.01014985263059953, "socrates": 0.014767948437162282}, {"ibn_sina": 0.017325430669877205, "einstein": 0.004115150483890806, "darwin": 0.013493058972749702, "sagan": 0.007222314020844984, "hypatia": 0.9459364631027943, "socrates": 0.011907582749842894}, {"einstein": 0.012628313596677996, "sagan": 0.005603780189386187, "socrates": 0.030293749491539623, "ibn_sina": 0.04994594915596257, "darwin": 0.016215075628037933, "hypatia": 0.8853131319383957}, {"socrates": 0.012140125213380253, "ibn_sina": 0.008343777902336211, "hypatia": 0.9644095990952897, "einstein": 0.0061044383178087995, "sagan": 0.005734589096888705, "darwin": 0.003267470374296399}], "prompt_tokens": 368, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.005584962174309691, "einstein": 0.0033874507920541607, "hypatia": 0.9392462198225568, "ibn_sina": 0.032139238273673335, "sagan": 0.009208045932840025, "socrates": 0.010434083004565957}, "confidence": 0.8278979508740782, "choice": "hypatia"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 193.8830310027697, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 368, "compiled_prompt_sha256": "a85f301fbcbe5359d5137e8ce9e30a12f45037248d0741358fb7ff25b1b3eafd", "compiled_ids_sha256": "c6dbc71f4fcb0e3fc82bd7695103d89d45b3a9f018a7f14f1ed0ba8bed8dbf04", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5, 15.0, 20.625, 17.25, 16.0, 16.125], "runtime_probabilities": [0.005584961734712124, 0.0033874509390443563, 0.9392462372779846, 0.032139237970113754, 0.009208046831190586, 0.010434083640575409], "decide_seconds": 0.18954447400028585, "probabilities_f64_minus_runtime_max_abs": 1.7455427769874632e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.011878897974596113, "hypatia": 0.9436577491595973, "sagan": 0.006768392042655814, "socrates": 0.02514762720956024, "darwin": 0.010483090668699494, "einstein": 0.002064242944891072}, "confidence": 0.835715453710377, "choice": "hypatia"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 194.0143000028911, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 368, "compiled_prompt_sha256": "b61c5ac999929626f974063e57c72670796723e26bac9fb12e40f33023fb5628", "compiled_ids_sha256": "8cedba15c25a3ae8f2644bdcb4b33b7308817bd310b112521064579da125c625", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 20.625, 15.6875, 17.0, 16.125, 14.5], "runtime_probabilities": [0.011878898367285728, 0.9436577558517456, 0.0067683919332921505, 0.025147628039121628, 0.010483091697096825, 0.002064242959022522], "decide_seconds": 0.18976699199993163, "probabilities_f64_minus_runtime_max_abs": 6.692148279974219e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.9136606169296246, "darwin": 0.006975884894319403, "sagan": 0.008957213508194192, "ibn_sina": 0.04548848360009997, "einstein": 0.01014985263059953, "socrates": 0.014767948437162282}, "confidence": 0.7718509509509576, "choice": "hypatia"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 193.9132109982893, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 368, "compiled_prompt_sha256": "081929d17758424a25168221bbd1848ce541fe1930acd1433380e80eb2c816a5", "compiled_ids_sha256": "dcc783e8b58640f809027b45c0b3df1e3a26c539c0d3ac436a4aa6ec9f9a1260", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.75, 15.875, 16.125, 17.75, 16.25, 16.625], "runtime_probabilities": [0.9136606454849243, 0.006975884549319744, 0.008957214653491974, 0.04548848420381546, 0.010149853304028511, 0.014767948538064957], "decide_seconds": 0.1895175939971523, "probabilities_f64_minus_runtime_max_abs": 2.855529968037729e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.017325430669877205, "einstein": 0.004115150483890806, "darwin": 0.013493058972749702, "sagan": 0.007222314020844984, "hypatia": 0.9459364631027943, "socrates": 0.011907582749842894}, "confidence": 0.837083218819121, "choice": "hypatia"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 193.89888099976815, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 368, "compiled_prompt_sha256": "02542a5ccc36f918849687042233556561fb2d6c471f687f7a179dbb318c5df6", "compiled_ids_sha256": "5d8fe48e10ed3cd7d04f6fe7c9bc2336ad464c766699d896f7a2b8e368b4a1a2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 14.9375, 16.125, 15.5, 20.375, 16.0], "runtime_probabilities": [0.01732543110847473, 0.004115150310099125, 0.013493058271706104, 0.007222313433885574, 0.9459364414215088, 0.011907583102583885], "decide_seconds": 0.18968572299854713, "probabilities_f64_minus_runtime_max_abs": 2.1681285544694617e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.012628313596677996, "sagan": 0.005603780189386187, "socrates": 0.030293749491539623, "ibn_sina": 0.04994594915596257, "darwin": 0.016215075628037933, "hypatia": 0.8853131319383957}, "confidence": 0.7128243690874736, "choice": "hypatia"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 194.29636800123262, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 368, "compiled_prompt_sha256": "fbab03d13c25e0f115e70de830408c8fb2989d1e6e51aee8b7ea316b12426d1d", "compiled_ids_sha256": "d0a7b7e25dce6d1359fbebcd1b4ae2af4c2c6e0ff3365d78e4e438a0f0168f70", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 15.4375, 17.125, 17.625, 16.5, 20.5], "runtime_probabilities": [0.012628312222659588, 0.005603779572993517, 0.030293747782707214, 0.049945950508117676, 0.016215074807405472, 0.885313093662262], "decide_seconds": 0.19003727000017534, "probabilities_f64_minus_runtime_max_abs": 3.8276133706993676e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.012140125213380253, "ibn_sina": 0.008343777902336211, "hypatia": 0.9644095990952897, "einstein": 0.0061044383178087995, "sagan": 0.005734589096888705, "darwin": 0.003267470374296399}, "confidence": 0.8839897868208577, "choice": "hypatia"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 193.9986700017471, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 368, "compiled_prompt_sha256": "ceffa6eefc3da573362e5e4396ae6d455b2ab4700f3b030698f74aa75fc83d91", "compiled_ids_sha256": "38f951a9aa964795d4df957a23f7e6e3286f5fb42a4d79c7acee8884325aaa48", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 16.0, 20.75, 15.6875, 15.625, 15.0625], "runtime_probabilities": [0.012140125967562199, 0.008343778550624847, 0.9644096493721008, 0.006104438565671444, 0.005734589416533709, 0.003267470747232437], "decide_seconds": 0.1897730419987056, "probabilities_f64_minus_runtime_max_abs": 5.027681115876703e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:18Z"} {"id": "c-hypatia-01-main-01", "uid": "c041", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "prompt_sha256_by_order": ["c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c"], "request_sha256": "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "compiled_prompt_sha256_by_order": ["937427227f5f99118cb41f68bee040c493d9bec3da24346374497e87fd4b0c16", "17b90d81647891d8db0520a988ad0cdae1e9f5bd1fed5dcd58e150aedbf9bc03", "0f3022a27181d5a0a95921e9f56219e94b652cc915de14ff442d6bba8f27fe29", "58f0b52eb9761ac361ed3c74d0385ea3ec690a657f7e3ee4907d149256845dab", "3a18fbc6a84be5fcb500032684a23491003db5acc02f325e46b5f30aead3dee2", "dd944680d9992422a25d628a6e81a6b820f291626cc233ff3d8ba7afc2cc8b85"], "compiled_tokens_by_order": [297, 297, 297, 297, 297, 297], "choice": "ibn_sina", "probs": {"darwin": 0.021763625362932436, "einstein": 0.01694952847512447, "hypatia": 0.2651352345469373, "ibn_sina": 0.6360263637148198, "sagan": 0.046073595254879994, "socrates": 0.014051652645306023}, "confidence": 0.44528381125138294, "seconds": 0.17477186749965767, "seconds_all_orders": 1.048631204997946, "seconds_by_order": [0.1752933350035164, 0.17456826099805767, 0.17475416899833363, 0.17458641099801753, 0.17456989000129397, 0.17485913899872685], "argmax_stable": false, "spread": {"darwin": 0.08007338712606255, "einstein": 0.07424914146540566, "hypatia": 0.6529418221802525, "ibn_sina": 0.5049246379503938, "sagan": 0.5031372559173986, "socrates": 0.049218353244796266}, "spread_2dp": {"darwin": 0.08, "einstein": 0.06999999999999999, "hypatia": 0.65, "ibn_sina": 0.51, "sagan": 0.51, "socrates": 0.049999999999999996}, "choices_by_order": ["ibn_sina", "sagan", "hypatia", "hypatia", "ibn_sina", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.021763625362932436, "einstein": 0.01694952847512447, "hypatia": 0.2651352345469373, "ibn_sina": 0.6360263637148198, "sagan": 0.046073595254879994, "socrates": 0.014051652645306023}, {"ibn_sina": 0.189620462684024, "hypatia": 0.10149651973268256, "sagan": 0.5154418580179789, "socrates": 0.06156075107200027, "darwin": 0.08957036428720863, "einstein": 0.04231004420610564}, {"hypatia": 0.594690030696884, "darwin": 0.015847952513001494, "sagan": 0.038017261743403065, "ibn_sina": 0.24790368727897721, "einstein": 0.09119866994053014, "socrates": 0.012342397827204003}, {"ibn_sina": 0.20148302044237454, "einstein": 0.024063715186776937, "darwin": 0.04495693861849639, "sagan": 0.09517383980221186, "hypatia": 0.6206113943374981, "socrates": 0.013711091612642173}, {"einstein": 0.055145400749029856, "sagan": 0.012304602100580373, "socrates": 0.029517205981177918, "ibn_sina": 0.5232049948969228, "darwin": 0.06248792555251339, "hypatia": 0.31733987071977565}, {"socrates": 0.0292527490655519, "ibn_sina": 0.131101725764426, "hypatia": 0.7544383419129351, "einstein": 0.04256249874435863, "sagan": 0.033147707351582185, "darwin": 0.009496977161146083}], "prompt_tokens": 297, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.021763625362932436, "einstein": 0.01694952847512447, "hypatia": 0.2651352345469373, "ibn_sina": 0.6360263637148198, "sagan": 0.046073595254879994, "socrates": 0.014051652645306023}, "confidence": 0.44528381125138294, "choice": "ibn_sina"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 175.26845500105992, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 297, "compiled_prompt_sha256": "937427227f5f99118cb41f68bee040c493d9bec3da24346374497e87fd4b0c16", "compiled_ids_sha256": "f2a92c82350a676ec89ba5633141c479e87276520dc9842cd8d2d1356c426518", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 16.125, 18.875, 19.75, 17.125, 15.9375], "runtime_probabilities": [0.021763626486063004, 0.016949528828263283, 0.2651352286338806, 0.6360263824462891, 0.04607360064983368, 0.014051652513444424], "decide_seconds": 0.17157687399958377, "probabilities_f64_minus_runtime_max_abs": 1.8731469264565703e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.189620462684024, "hypatia": 0.10149651973268256, "sagan": 0.5154418580179789, "socrates": 0.06156075107200027, "darwin": 0.08957036428720863, "einstein": 0.04231004420610564}, "confidence": 0.2127163578829241, "choice": "sagan"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 174.54549099784344, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 297, "compiled_prompt_sha256": "17b90d81647891d8db0520a988ad0cdae1e9f5bd1fed5dcd58e150aedbf9bc03", "compiled_ids_sha256": "ad9bc07ae62caa648cf4adc6a4b8d09b9940231b74b8f8c2faf04a6a55170a49", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.625, 18.0, 19.625, 17.5, 17.875, 17.125], "runtime_probabilities": [0.1896204799413681, 0.10149651020765305, 0.5154418349266052, 0.06156075373291969, 0.08957035839557648, 0.042310044169425964], "decide_seconds": 0.17090489999827696, "probabilities_f64_minus_runtime_max_abs": 2.3091373702044393e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.594690030696884, "darwin": 0.015847952513001494, "sagan": 0.038017261743403065, "ibn_sina": 0.24790368727897721, "einstein": 0.09119866994053014, "socrates": 0.012342397827204003}, "confidence": 0.3763385616613766, "choice": "hypatia"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 174.73237900048844, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 297, "compiled_prompt_sha256": "0f3022a27181d5a0a95921e9f56219e94b652cc915de14ff442d6bba8f27fe29", "compiled_ids_sha256": "e6d851f1dddf7852b037b497b70026a76a6254e6f7ec9a64c0052f10a335a0de", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.875, 16.25, 17.125, 19.0, 18.0, 16.0], "runtime_probabilities": [0.5946900248527527, 0.01584795117378235, 0.038017258048057556, 0.24790367484092712, 0.09119866788387299, 0.01234239712357521], "decide_seconds": 0.17097865899995668, "probabilities_f64_minus_runtime_max_abs": 1.2438050089880193e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.20148302044237454, "einstein": 0.024063715186776937, "darwin": 0.04495693861849639, "sagan": 0.09517383980221186, "hypatia": 0.6206113943374981, "socrates": 0.013711091612642173}, "confidence": 0.36896547277166214, "choice": "hypatia"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 174.56332100118743, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 297, "compiled_prompt_sha256": "58f0b52eb9761ac361ed3c74d0385ea3ec690a657f7e3ee4907d149256845dab", "compiled_ids_sha256": "41784a82da1e97accdcff700414b1231236af276fc5ece9e41e0ae6ce5b977a8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.625, 16.5, 17.125, 17.875, 19.75, 15.9375], "runtime_probabilities": [0.20148301124572754, 0.02406371384859085, 0.04495694115757942, 0.09517384320497513, 0.6206113696098328, 0.013711091130971909], "decide_seconds": 0.17083969000304933, "probabilities_f64_minus_runtime_max_abs": 2.4727665359769446e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.055145400749029856, "sagan": 0.012304602100580373, "socrates": 0.029517205981177918, "ibn_sina": 0.5232049948969228, "darwin": 0.06248792555251339, "hypatia": 0.31733987071977565}, "confidence": 0.33343726851054867, "choice": "ibn_sina"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 174.54749099852052, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 297, "compiled_prompt_sha256": "3a18fbc6a84be5fcb500032684a23491003db5acc02f325e46b5f30aead3dee2", "compiled_ids_sha256": "a8a433667a748fb07b94519749f67325473a0289572d2acb178a434eeed3577a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.25, 15.75, 16.625, 19.5, 17.375, 19.0], "runtime_probabilities": [0.055145397782325745, 0.012304602190852165, 0.02951720729470253, 0.5232049822807312, 0.062487926334142685, 0.31733986735343933], "decide_seconds": 0.17096078900067369, "probabilities_f64_minus_runtime_max_abs": 1.2616191646941388e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0292527490655519, "ibn_sina": 0.131101725764426, "hypatia": 0.7544383419129351, "einstein": 0.04256249874435863, "sagan": 0.033147707351582185, "darwin": 0.009496977161146083}, "confidence": 0.5123317395138383, "choice": "hypatia"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 174.8372889996972, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 297, "compiled_prompt_sha256": "dd944680d9992422a25d628a6e81a6b820f291626cc233ff3d8ba7afc2cc8b85", "compiled_ids_sha256": "0a7ed79ec519fae4996b7c78e47c1a5a5e9f9ca8dd751c48cd6d294f6d1cbee1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 18.375, 20.125, 17.25, 17.0, 15.75], "runtime_probabilities": [0.029252750799059868, 0.13110172748565674, 0.7544384002685547, 0.04256250336766243, 0.03314771130681038, 0.009496977552771568], "decide_seconds": 0.17121461699935026, "probabilities_f64_minus_runtime_max_abs": 5.8355619558803085e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:19Z"} {"id": "c-hypatia-01-main-01", "uid": "c042", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "prompt_sha256_by_order": ["79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60"], "request_sha256": "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "compiled_prompt_sha256_by_order": ["9e620d83c4565136473c23f396206defac1ffa16bd69dabc4a8fb76aa81893ae", "4ca5f41c1e88468acb0f861f5f2429161ef8d500bc13ba5dd73a19e226767655", "3cbbcc1ebea5302b168d6eccfe70024bcfe034f98c60ecb65abaf633d9aed5d2", "5c791f350bba1210c9a65b72356a61dc168f95d3e7f029aebef1e44ef4d75cc1", "a71462eb6cbb64abcd306c0d7907f55029c8e139b3bbc780467b01b369535509", "65fd84f15462d8af5cf515f76ed6c8b4f2a16c5d42201402c3045864dd35bf66"], "compiled_tokens_by_order": [388, 388, 388, 388, 388, 388], "choice": "hypatia", "probs": {"darwin": 0.017496788392115874, "einstein": 0.0041558515104352664, "hypatia": 0.6565621405962685, "ibn_sina": 0.2415357133768812, "sagan": 0.04756120194268173, "socrates": 0.032688304181617514}, "confidence": 0.4588271533024708, "seconds": 0.21063004700044985, "seconds_all_orders": 1.263780282002699, "seconds_by_order": [0.21132245500120916, 0.2103588520003541, 0.21080254900152795, 0.21019403399986913, 0.2105223010003101, 0.21058009099942865], "argmax_stable": false, "spread": {"darwin": 0.05405152248525247, "einstein": 0.20018276507360291, "hypatia": 0.6183048101523134, "ibn_sina": 0.5486695551583154, "sagan": 0.045961124855781396, "socrates": 0.4505895924005464}, "spread_2dp": {"darwin": 0.049999999999999996, "einstein": 0.2, "hypatia": 0.62, "ibn_sina": 0.5499999999999999, "sagan": 0.05, "socrates": 0.44999999999999996}, "choices_by_order": ["hypatia", "socrates", "hypatia", "hypatia", "ibn_sina", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.017496788392115874, "einstein": 0.0041558515104352664, "hypatia": 0.6565621405962685, "ibn_sina": 0.2415357133768812, "sagan": 0.04756120194268173, "socrates": 0.032688304181617514}, {"ibn_sina": 0.08398110587151904, "hypatia": 0.2586800172962219, "sagan": 0.050937115547648945, "socrates": 0.4832778965821639, "darwin": 0.05771931368650459, "einstein": 0.06540455101594157}, {"hypatia": 0.6294072496867558, "darwin": 0.0065684540669406315, "sagan": 0.027654224551578413, "ibn_sina": 0.09652272790767005, "einstein": 0.20433861658403818, "socrates": 0.035508727203016836}, {"ibn_sina": 0.05485403445600648, "einstein": 0.010801399816440817, "darwin": 0.02286656359084535, "sagan": 0.015715943998780003, "hypatia": 0.8580614683578835, "socrates": 0.037700589780043835}, {"einstein": 0.04166345996488244, "sagan": 0.004975990691867547, "socrates": 0.07783759065203086, "ibn_sina": 0.575146323933456, "darwin": 0.060619976552193096, "hypatia": 0.23975665820557007}, {"socrates": 0.08155418928609504, "ibn_sina": 0.026476768775140644, "hypatia": 0.7737646042948505, "einstein": 0.06351446647876441, "sagan": 0.04365281185898336, "darwin": 0.011037159306165997}], "prompt_tokens": 388, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.017496788392115874, "einstein": 0.0041558515104352664, "hypatia": 0.6565621405962685, "ibn_sina": 0.2415357133768812, "sagan": 0.04756120194268173, "socrates": 0.032688304181617514}, "confidence": 0.4588271533024708, "choice": "hypatia"}}, "usage": {"input_tokens": 388, "output_tokens": 1}, "latency_ms": 211.29468499930226, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 388, "compiled_prompt_sha256": "9e620d83c4565136473c23f396206defac1ffa16bd69dabc4a8fb76aa81893ae", "compiled_ids_sha256": "5cdc4054cab42c152b7033db49be91482544fff96b21837effc480511601a69b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 15.0625, 20.125, 19.125, 17.5, 17.125], "runtime_probabilities": [0.017496787011623383, 0.004155851434916258, 0.6565621495246887, 0.2415357381105423, 0.04756120219826698, 0.03268830105662346], "decide_seconds": 0.20681110000077751, "probabilities_f64_minus_runtime_max_abs": 2.473366109145836e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.08398110587151904, "hypatia": 0.2586800172962219, "sagan": 0.050937115547648945, "socrates": 0.4832778965821639, "darwin": 0.05771931368650459, "einstein": 0.06540455101594157}, "confidence": 0.2164815873257797, "choice": "socrates"}}, "usage": {"input_tokens": 388, "output_tokens": 1}, "latency_ms": 210.3340220019163, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 388, "compiled_prompt_sha256": "4ca5f41c1e88468acb0f861f5f2429161ef8d500bc13ba5dd73a19e226767655", "compiled_ids_sha256": "62c8a0a4a87a35dd1b0e0a2142b795c557227c4674a4b0759129b0090c13b730", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.875, 19.0, 17.375, 19.625, 17.5, 17.625], "runtime_probabilities": [0.08398109674453735, 0.2586800158023834, 0.05093711242079735, 0.4832778871059418, 0.057719312608242035, 0.06540454924106598], "decide_seconds": 0.2059179870011576, "probabilities_f64_minus_runtime_max_abs": 9.476222140314405e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.6294072496867558, "darwin": 0.0065684540669406315, "sagan": 0.027654224551578413, "ibn_sina": 0.09652272790767005, "einstein": 0.20433861658403818, "socrates": 0.035508727203016836}, "confidence": 0.3903682568083441, "choice": "hypatia"}}, "usage": {"input_tokens": 388, "output_tokens": 1}, "latency_ms": 210.77714899729472, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 388, "compiled_prompt_sha256": "3cbbcc1ebea5302b168d6eccfe70024bcfe034f98c60ecb65abaf633d9aed5d2", "compiled_ids_sha256": "cb010ed0292e1d26c5d8f62251685a997452f9d2a0e088f9600cc7333ca82088", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.875, 15.3125, 16.75, 18.0, 18.75, 17.0], "runtime_probabilities": [0.6294072270393372, 0.006568454205989838, 0.027654225006699562, 0.0965227261185646, 0.20433861017227173, 0.03550872951745987], "decide_seconds": 0.20621420400129864, "probabilities_f64_minus_runtime_max_abs": 2.2647418607135705e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.05485403445600648, "einstein": 0.010801399816440817, "darwin": 0.02286656359084535, "sagan": 0.015715943998780003, "hypatia": 0.8580614683578835, "socrates": 0.037700589780043835}, "confidence": 0.6568994431446101, "choice": "hypatia"}}, "usage": {"input_tokens": 388, "output_tokens": 1}, "latency_ms": 210.16867400248884, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 388, "compiled_prompt_sha256": "5c791f350bba1210c9a65b72356a61dc168f95d3e7f029aebef1e44ef4d75cc1", "compiled_ids_sha256": "0f8d36ea8a5c218b7d2162d67f980a8151b2c6399f0597c968d3841fb62a5132", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 16.0, 16.75, 16.375, 20.375, 17.25], "runtime_probabilities": [0.05485403165221214, 0.010801400057971478, 0.022866563871502876, 0.0157159436494112, 0.8580614924430847, 0.0377005897462368], "decide_seconds": 0.205791518001206, "probabilities_f64_minus_runtime_max_abs": 2.4085201166812453e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.04166345996488244, "sagan": 0.004975990691867547, "socrates": 0.07783759065203086, "ibn_sina": 0.575146323933456, "darwin": 0.060619976552193096, "hypatia": 0.23975665820557007}, "confidence": 0.33696985715678407, "choice": "ibn_sina"}}, "usage": {"input_tokens": 388, "output_tokens": 1}, "latency_ms": 210.4978810020839, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 388, "compiled_prompt_sha256": "a71462eb6cbb64abcd306c0d7907f55029c8e139b3bbc780467b01b369535509", "compiled_ids_sha256": "be59331babc1741f9c6a119ab84993967e55001c780438b5c00f15bbc2552ce0", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 15.25, 18.0, 20.0, 17.75, 19.125], "runtime_probabilities": [0.041663460433483124, 0.004975990857928991, 0.07783758640289307, 0.5751463174819946, 0.06061997264623642, 0.23975665867328644], "decide_seconds": 0.2060950449995289, "probabilities_f64_minus_runtime_max_abs": 6.451461365131195e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.08155418928609504, "ibn_sina": 0.026476768775140644, "hypatia": 0.7737646042948505, "einstein": 0.06351446647876441, "sagan": 0.04365281185898336, "darwin": 0.011037159306165997}, "confidence": 0.519723744677496, "choice": "hypatia"}}, "usage": {"input_tokens": 388, "output_tokens": 1}, "latency_ms": 210.5549109983258, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 388, "compiled_prompt_sha256": "65fd84f15462d8af5cf515f76ed6c8b4f2a16c5d42201402c3045864dd35bf66", "compiled_ids_sha256": "72dd8f4c1e51c227ef6c4f5655c588ade075dd15eacbbc13317c7afd4a73b761", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 17.0, 20.375, 17.875, 17.5, 16.125], "runtime_probabilities": [0.08155418932437897, 0.026476768776774406, 0.7737646102905273, 0.06351446360349655, 0.043652813881635666, 0.011037158779799938], "decide_seconds": 0.20618524499877822, "probabilities_f64_minus_runtime_max_abs": 5.995676843895126e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:20Z"} {"id": "c-hypatia-01-main-01", "uid": "c043", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "prompt_sha256_by_order": ["1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568"], "request_sha256": "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "compiled_prompt_sha256_by_order": ["96b3447c5f6b1e9a91ad93d8847b48d6895167515fc8658b556b546cd722af75", "727f593328d2a66d56ff3b237949178a2b87a72508a3a53f2e7fe994cc222f35", "d5b11cdefbb4228224e0a6f9b37c43c8ade2eccc794c32841a45140b4935c41d", "3eaa0693bfbe41e644e724319a2e0700d21d42a1e292dfee99cd7bdda1af2c9e", "886ceda06e84376619e086e452d366d1d5adab0db0db140effd8d3e228b04c96", "8228536bbf2e08d2ec210e318ca04e418945ededf94ccd1d71c6e258bcd2735c"], "compiled_tokens_by_order": [339, 339, 339, 339, 339, 339], "choice": "ibn_sina", "probs": {"darwin": 0.08749113342330919, "einstein": 0.09914044249568611, "hypatia": 0.046830629058906045, "ibn_sina": 0.4443166375613959, "sagan": 0.20988031841035543, "socrates": 0.11234083905034736}, "confidence": 0.15203496195078414, "seconds": 0.18558489016625876, "seconds_all_orders": 1.1135093409975525, "seconds_by_order": [0.18587933299932047, 0.18561385499924654, 0.18561692500225035, 0.18567314499887289, 0.18539981600042665, 0.1853262669974356], "argmax_stable": false, "spread": {"darwin": 0.07903427097503257, "einstein": 0.3529592484398446, "hypatia": 0.11361377391574778, "ibn_sina": 0.5327649446961057, "sagan": 0.18128749687171816, "socrates": 0.5421884030058419}, "spread_2dp": {"darwin": 0.08, "einstein": 0.35000000000000003, "hypatia": 0.11, "ibn_sina": 0.54, "sagan": 0.18, "socrates": 0.54}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "einstein", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.08749113342330919, "einstein": 0.09914044249568611, "hypatia": 0.046830629058906045, "ibn_sina": 0.4443166375613959, "sagan": 0.20988031841035543, "socrates": 0.11234083905034736}, {"ibn_sina": 0.5068198886016086, "hypatia": 0.036713909590426926, "sagan": 0.028592821538637254, "socrates": 0.34833187571392055, "darwin": 0.032399911495322634, "einstein": 0.047141593060084}, {"hypatia": 0.13353362186554463, "darwin": 0.016976894886945308, "sagan": 0.07147539719506608, "ibn_sina": 0.4660781366627762, "einstein": 0.22015972272335113, "socrates": 0.09177622666631656}, {"ibn_sina": 0.6570376311956542, "einstein": 0.06925125683598615, "darwin": 0.03706752666076544, "sagan": 0.06925125683598615, "hypatia": 0.07847195455663118, "socrates": 0.08892037391497692}, {"einstein": 0.4001008414999286, "sagan": 0.07878452695035867, "socrates": 0.24267342734653147, "ibn_sina": 0.18899425524810037, "darwin": 0.0695271010052841, "hypatia": 0.01991984794979684}, {"socrates": 0.6311087769208188, "ibn_sina": 0.12427268649954844, "hypatia": 0.08541128507769125, "einstein": 0.07537519452683239, "sagan": 0.07537519452683239, "darwin": 0.008456862448276625}], "prompt_tokens": 339, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.08749113342330919, "einstein": 0.09914044249568611, "hypatia": 0.046830629058906045, "ibn_sina": 0.4443166375613959, "sagan": 0.20988031841035543, "socrates": 0.11234083905034736}, "confidence": 0.15203496195078414, "choice": "ibn_sina"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 185.85240300308215, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "96b3447c5f6b1e9a91ad93d8847b48d6895167515fc8658b556b546cd722af75", "compiled_ids_sha256": "a5315fe99a572917cd0eab60dbea5014cc09871c3719dbb84d88bd7901126ebf", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 17.5, 16.75, 19.0, 18.25, 17.625], "runtime_probabilities": [0.08749112486839294, 0.09914044290781021, 0.04683062434196472, 0.4443166255950928, 0.20988032221794128, 0.11234083026647568], "decide_seconds": 0.1817059960012557, "probabilities_f64_minus_runtime_max_abs": 1.196630311506297e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.5068198886016086, "hypatia": 0.036713909590426926, "sagan": 0.028592821538637254, "socrates": 0.34833187571392055, "darwin": 0.032399911495322634, "einstein": 0.047141593060084}, "confidence": 0.33592456109799806, "choice": "ibn_sina"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 185.58972499886295, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "727f593328d2a66d56ff3b237949178a2b87a72508a3a53f2e7fe994cc222f35", "compiled_ids_sha256": "fe54e2af12285f2a723f14ddc1a2bd915bb9e4b9d0aa5b734019194b4aac6475", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.25, 16.625, 16.375, 18.875, 16.5, 16.875], "runtime_probabilities": [0.5068199038505554, 0.03671390935778618, 0.028592823073267937, 0.34833186864852905, 0.03239991143345833, 0.04714158922433853], "decide_seconds": 0.1815008370031137, "probabilities_f64_minus_runtime_max_abs": 1.524894677196187e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.13353362186554463, "darwin": 0.016976894886945308, "sagan": 0.07147539719506608, "ibn_sina": 0.4660781366627762, "einstein": 0.22015972272335113, "socrates": 0.09177622666631656}, "confidence": 0.1992075219507361, "choice": "ibn_sina"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 185.59418500080938, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "d5b11cdefbb4228224e0a6f9b37c43c8ade2eccc794c32841a45140b4935c41d", "compiled_ids_sha256": "23fbfc222e8036a868893934c7aa06b92cfb2f659933da08dad21e131ac6da46", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.0, 15.9375, 17.375, 19.25, 18.5, 17.625], "runtime_probabilities": [0.13353361189365387, 0.016976894810795784, 0.07147539407014847, 0.46607810258865356, 0.22015972435474396, 0.09177621454000473], "decide_seconds": 0.1816436760018405, "probabilities_f64_minus_runtime_max_abs": 3.407412263456422e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.6570376311956542, "einstein": 0.06925125683598615, "darwin": 0.03706752666076544, "sagan": 0.06925125683598615, "hypatia": 0.07847195455663118, "socrates": 0.08892037391497692}, "confidence": 0.3398626035964022, "choice": "ibn_sina"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 185.64979499933543, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "3eaa0693bfbe41e644e724319a2e0700d21d42a1e292dfee99cd7bdda1af2c9e", "compiled_ids_sha256": "232e1dc5fe06f394111fd1f330f33a99b1e2babef926933a554298012f04b014", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.375, 17.125, 16.5, 17.125, 17.25, 17.375], "runtime_probabilities": [0.6570376753807068, 0.06925125420093536, 0.03706752881407738, 0.06925125420093536, 0.07847195863723755, 0.08892037719488144], "decide_seconds": 0.1816706850004266, "probabilities_f64_minus_runtime_max_abs": 4.418505261671868e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.4001008414999286, "sagan": 0.07878452695035867, "socrates": 0.24267342734653147, "ibn_sina": 0.18899425524810037, "darwin": 0.0695271010052841, "hypatia": 0.01991984794979684}, "confidence": 0.16920886007607883, "choice": "einstein"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 185.3773570001067, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "886ceda06e84376619e086e452d366d1d5adab0db0db140effd8d3e228b04c96", "compiled_ids_sha256": "4e2f2e110205e1684b6808fe1ab6b3a09caa33a2ca0a3209142a652ad8eef405", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.125, 17.5, 18.625, 18.375, 17.375, 16.125], "runtime_probabilities": [0.40010085701942444, 0.0787845253944397, 0.24267342686653137, 0.18899425864219666, 0.06952709704637527, 0.019919848069548607], "decide_seconds": 0.18144381800084375, "probabilities_f64_minus_runtime_max_abs": 1.5519495855098597e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.6311087769208188, "ibn_sina": 0.12427268649954844, "hypatia": 0.08541128507769125, "einstein": 0.07537519452683239, "sagan": 0.07537519452683239, "darwin": 0.008456862448276625}, "confidence": 0.33592787126278645, "choice": "socrates"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 185.3029370031436, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "8228536bbf2e08d2ec210e318ca04e418945ededf94ccd1d71c6e258bcd2735c", "compiled_ids_sha256": "48b0351d721ebeacce205e0e28dcd105814af68abcf16e166271126420fa4fa4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.75, 18.125, 17.75, 17.625, 17.625, 15.4375], "runtime_probabilities": [0.6311087608337402, 0.1242726743221283, 0.08541128039360046, 0.07537519186735153, 0.07537519186735153, 0.008456862531602383], "decide_seconds": 0.1812932990033005, "probabilities_f64_minus_runtime_max_abs": 1.608707855815794e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:22Z"} {"id": "c-hypatia-01-main-01", "uid": "c044", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "prompt_sha256_by_order": ["550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0"], "request_sha256": "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "compiled_prompt_sha256_by_order": ["12910ec5c352e9fa1330272cb09c7a1e3ca9ef566bd28df789137a8794a156bc", "807d8a79dbcf0eff03407ca148654a28f33c941d1b260bbbe9deb3edbed4c903", "3775085ccc872c9b384b145b06b80f3a2a0cb2deb35c9b642e3d16edf2c1a06b", "dddb54acc96be195aa3e9ab1c24a8384413f94d3b46ba7004698c660b361c35c", "e414d533753b06ef4a940fdcc038fbe8a9bff466fd2a6493da0716b9f484077f", "d2b15f3e99a322a75aff55d2228be720fa0786b4ffb917d8ca30879880ae7d96"], "compiled_tokens_by_order": [275, 275, 275, 275, 275, 275], "choice": "einstein", "probs": {"darwin": 0.004232237814236728, "einstein": 0.9728509527612262, "hypatia": 0.003096371330314564, "ibn_sina": 0.007427808685239645, "sagan": 0.00841680992135564, "socrates": 0.003975819487627134}, "confidence": 0.9071310940527754, "seconds": 0.17279939949973291, "seconds_all_orders": 1.0367963969983975, "seconds_by_order": [0.17299566300061997, 0.17290120399775333, 0.17283699399922625, 0.17268078599954606, 0.17272686499927659, 0.17265488500197534], "argmax_stable": true, "spread": {"darwin": 0.0233669514230103, "einstein": 0.40497033903239965, "hypatia": 0.19231511585677838, "ibn_sina": 0.16192460083284974, "sagan": 0.08423197504983135, "socrates": 0.15904488581105902}, "spread_2dp": {"darwin": 0.03, "einstein": 0.4, "hypatia": 0.2, "ibn_sina": 0.16, "sagan": 0.08, "socrates": 0.16}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.004232237814236728, "einstein": 0.9728509527612262, "hypatia": 0.003096371330314564, "ibn_sina": 0.007427808685239645, "sagan": 0.00841680992135564, "socrates": 0.003975819487627134}, {"ibn_sina": 0.1693524095180894, "hypatia": 0.02291935630893345, "sagan": 0.09064781264178558, "socrates": 0.029429036034793094, "darwin": 0.017849612640889915, "einstein": 0.6698017728555086}, {"hypatia": 0.19541148718709295, "darwin": 0.0052075397426908235, "sagan": 0.029967322138694807, "ibn_sina": 0.08145962721719253, "einstein": 0.6820531081106147, "socrates": 0.005900915603714125}, {"ibn_sina": 0.07342537301348975, "einstein": 0.8945041632992388, "darwin": 0.003434150394615262, "sagan": 0.006415837591954234, "hypatia": 0.01856484163454349, "socrates": 0.003655634066158526}, {"einstein": 0.7832430051680611, "sagan": 0.018420109939039288, "socrates": 0.03899537304695347, "ibn_sina": 0.09354503698136715, "darwin": 0.026801101817625562, "hypatia": 0.03899537304695347}, {"socrates": 0.16270051987721754, "ibn_sina": 0.02495093283142332, "hypatia": 0.16270051987721754, "einstein": 0.5678806137288266, "sagan": 0.0768542837035722, "darwin": 0.0049131299817428906}], "prompt_tokens": 275, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.004232237814236728, "einstein": 0.9728509527612262, "hypatia": 0.003096371330314564, "ibn_sina": 0.007427808685239645, "sagan": 0.00841680992135564, "socrates": 0.003975819487627134}, "confidence": 0.9071310940527754, "choice": "einstein"}}, "usage": {"input_tokens": 275, "output_tokens": 1}, "latency_ms": 172.9707740014419, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 275, "compiled_prompt_sha256": "12910ec5c352e9fa1330272cb09c7a1e3ca9ef566bd28df789137a8794a156bc", "compiled_ids_sha256": "bf822f6c8b673ada113561218d839770895abe054bccedd183916316a0e84576", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.4375, 20.875, 15.125, 16.0, 16.125, 15.375], "runtime_probabilities": [0.004232238046824932, 0.9728509783744812, 0.0030963716562837362, 0.007427808362990618, 0.00841681007295847, 0.003975819796323776], "decide_seconds": 0.1694294910012104, "probabilities_f64_minus_runtime_max_abs": 2.561325496142075e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.1693524095180894, "hypatia": 0.02291935630893345, "sagan": 0.09064781264178558, "socrates": 0.029429036034793094, "darwin": 0.017849612640889915, "einstein": 0.6698017728555086}, "confidence": 0.4145687023605322, "choice": "einstein"}}, "usage": {"input_tokens": 275, "output_tokens": 1}, "latency_ms": 172.87921400202322, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 275, "compiled_prompt_sha256": "807d8a79dbcf0eff03407ca148654a28f33c941d1b260bbbe9deb3edbed4c903", "compiled_ids_sha256": "24d303cadcd22aa980d79c46d494051003717dd886114c4507fb2524aa82a045", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 16.125, 17.5, 16.375, 15.875, 19.5], "runtime_probabilities": [0.16935241222381592, 0.02291935496032238, 0.09064780920743942, 0.029429035261273384, 0.01784961298108101, 0.6698017716407776], "decide_seconds": 0.16943902099956176, "probabilities_f64_minus_runtime_max_abs": 3.434346154684853e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.19541148718709295, "darwin": 0.0052075397426908235, "sagan": 0.029967322138694807, "ibn_sina": 0.08145962721719253, "einstein": 0.6820531081106147, "socrates": 0.005900915603714125}, "confidence": 0.471425811794006, "choice": "einstein"}}, "usage": {"input_tokens": 275, "output_tokens": 1}, "latency_ms": 172.8146139976161, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 275, "compiled_prompt_sha256": "3775085ccc872c9b384b145b06b80f3a2a0cb2deb35c9b642e3d16edf2c1a06b", "compiled_ids_sha256": "73cdc93d43cefdf705b52bb4cf5773efb4c2b9e507e3baf50004a8f49463a85e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.375, 14.75, 16.5, 17.5, 19.625, 14.875], "runtime_probabilities": [0.19541148841381073, 0.005207539536058903, 0.029967322945594788, 0.08145962655544281, 0.6820530891418457, 0.005900915712118149], "decide_seconds": 0.16942752099930658, "probabilities_f64_minus_runtime_max_abs": 1.89687689999829e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.07342537301348975, "einstein": 0.8945041632992388, "darwin": 0.003434150394615262, "sagan": 0.006415837591954234, "hypatia": 0.01856484163454349, "socrates": 0.003655634066158526}, "confidence": 0.7556175535026249, "choice": "einstein"}}, "usage": {"input_tokens": 275, "output_tokens": 1}, "latency_ms": 172.6597960005165, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 275, "compiled_prompt_sha256": "dddb54acc96be195aa3e9ab1c24a8384413f94d3b46ba7004698c660b361c35c", "compiled_ids_sha256": "58d50768ea1738f96ec68f9834c8140601d064cdc382fff1ca4128c75457b6aa", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 20.25, 14.6875, 15.3125, 16.375, 14.75], "runtime_probabilities": [0.07342538237571716, 0.8945042490959167, 0.0034341507125645876, 0.006415837910026312, 0.018564842641353607, 0.003655634354799986], "decide_seconds": 0.16930095199859352, "probabilities_f64_minus_runtime_max_abs": 8.579667798436219e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.7832430051680611, "sagan": 0.018420109939039288, "socrates": 0.03899537304695347, "ibn_sina": 0.09354503698136715, "darwin": 0.026801101817625562, "hypatia": 0.03899537304695347}, "confidence": 0.533086406488535, "choice": "einstein"}}, "usage": {"input_tokens": 275, "output_tokens": 1}, "latency_ms": 172.70594500223524, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 275, "compiled_prompt_sha256": "e414d533753b06ef4a940fdcc038fbe8a9bff466fd2a6493da0716b9f484077f", "compiled_ids_sha256": "270b5d8a6c423fdc32fb16d3c6ace64acc0368b85487141cfc4f2f40250e5fb7", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.125, 16.375, 17.125, 18.0, 16.75, 17.125], "runtime_probabilities": [0.7832430005073547, 0.018420111387968063, 0.038995373994112015, 0.09354504197835922, 0.026801101863384247, 0.038995373994112015], "decide_seconds": 0.1693349009983649, "probabilities_f64_minus_runtime_max_abs": 4.996992070127604e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.16270051987721754, "ibn_sina": 0.02495093283142332, "hypatia": 0.16270051987721754, "einstein": 0.5678806137288266, "sagan": 0.0768542837035722, "darwin": 0.0049131299817428906}, "confidence": 0.3148561566742638, "choice": "einstein"}}, "usage": {"input_tokens": 275, "output_tokens": 1}, "latency_ms": 172.6331560021208, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 275, "compiled_prompt_sha256": "d2b15f3e99a322a75aff55d2228be720fa0786b4ffb917d8ca30879880ae7d96", "compiled_ids_sha256": "ddb453eb0bdf05edbcd3f6e05e44b2cc37bfd387ac0ba8e59a25ed59290ad899", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.25, 16.375, 18.25, 19.5, 17.5, 14.75], "runtime_probabilities": [0.16270053386688232, 0.024950934574007988, 0.16270053386688232, 0.5678806304931641, 0.07685428857803345, 0.004913130309432745], "decide_seconds": 0.16922572299881722, "probabilities_f64_minus_runtime_max_abs": 1.6764337473773594e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:23Z"} {"id": "c-hypatia-01-main-02", "uid": "c045", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "prompt_sha256_by_order": ["a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2"], "request_sha256": "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "compiled_prompt_sha256_by_order": ["53ea90fe5d10904b59900dd3288e3f69d354a8b53be8832d2ed7310f00a05a78", "7b8ba837eec9ce754951db3caac4fd2da2ca873b5a41214f681f3059bf74da7a", "4f5ec06b6ec67c809734ca47b5b836ce712df5bf1d4eed101c7b9f244a430910", "dc1f71e89449ba9f6ec5ec75538f0ac800608cc1bf7eabab422b36013ac66370", "56e7f69d049c3c61a0c93373197985ba905699c9fdca194d77cb1dc097f03228", "5af5f17bf2c762b293eb16ba8999f89bbfdbe8db14a42dd8732b5bafceca03b6"], "compiled_tokens_by_order": [349, 349, 349, 349, 349, 349], "choice": "ibn_sina", "probs": {"darwin": 0.023399087786020182, "einstein": 0.1728971723147061, "hypatia": 0.19591816334803772, "ibn_sina": 0.46998324169501793, "sagan": 0.05613149359125109, "socrates": 0.08167084126496693}, "confidence": 0.20090587482926714, "seconds": 0.18658885733384523, "seconds_all_orders": 1.1195331440030714, "seconds_by_order": [0.18640314900039812, 0.1870000839990098, 0.1865747280025971, 0.1865838170015195, 0.18652329700125847, 0.18644806899828836], "argmax_stable": false, "spread": {"darwin": 0.03695674728898851, "einstein": 0.29759743726914406, "hypatia": 0.5107644293922998, "ibn_sina": 0.46845679772327664, "sagan": 0.04013255566309559, "socrates": 0.19655128125681434}, "spread_2dp": {"darwin": 0.04, "einstein": 0.3, "hypatia": 0.52, "ibn_sina": 0.4700000000000001, "sagan": 0.039999999999999994, "socrates": 0.19999999999999998}, "choices_by_order": ["ibn_sina", "einstein", "ibn_sina", "hypatia", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.023399087786020182, "einstein": 0.1728971723147061, "hypatia": 0.19591816334803772, "ibn_sina": 0.46998324169501793, "sagan": 0.05613149359125109, "socrates": 0.08167084126496693}, {"ibn_sina": 0.21183535383373767, "hypatia": 0.12848463691123604, "sagan": 0.028668797610616956, "socrates": 0.24004090350156362, "darwin": 0.04171285439087447, "einstein": 0.3492574537519712}, {"hypatia": 0.11821704987445025, "darwin": 0.008044771204273563, "sagan": 0.01599893792815549, "ibn_sina": 0.6802921515570143, "einstein": 0.13395746719135715, "socrates": 0.043489622244749275}, {"ibn_sina": 0.23152413124457064, "einstein": 0.05166001648282714, "darwin": 0.024402463900537054, "sagan": 0.027651614219912635, "hypatia": 0.5553975183997963, "socrates": 0.1093642557523561}, {"einstein": 0.200031627180323, "sagan": 0.01641959584183305, "socrates": 0.15578478788708286, "ibn_sina": 0.5437423372813665, "darwin": 0.03938856280189819, "hypatia": 0.044633089007496476}, {"socrates": 0.08430401554802439, "ibn_sina": 0.622927100249474, "hypatia": 0.0743980325965751, "einstein": 0.1784716023156828, "sagan": 0.03514314218835782, "darwin": 0.004756107101885958}], "prompt_tokens": 349, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.023399087786020182, "einstein": 0.1728971723147061, "hypatia": 0.19591816334803772, "ibn_sina": 0.46998324169501793, "sagan": 0.05613149359125109, "socrates": 0.08167084126496693}, "confidence": 0.20090587482926714, "choice": "ibn_sina"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 186.37693900018348, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 349, "compiled_prompt_sha256": "53ea90fe5d10904b59900dd3288e3f69d354a8b53be8832d2ed7310f00a05a78", "compiled_ids_sha256": "fc14aba23487d6da377ebf5043577838bb63cffc65b91089be20b810df6d2bad", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 18.5, 18.625, 19.5, 17.375, 17.75], "runtime_probabilities": [0.02339908853173256, 0.17289718985557556, 0.19591817259788513, 0.4699832499027252, 0.05613149702548981, 0.0816708356142044], "decide_seconds": 0.1820856320009625, "probabilities_f64_minus_runtime_max_abs": 1.7540869451382335e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.21183535383373767, "hypatia": 0.12848463691123604, "sagan": 0.028668797610616956, "socrates": 0.24004090350156362, "darwin": 0.04171285439087447, "einstein": 0.3492574537519712}, "confidence": 0.14236512905620302, "choice": "einstein"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 186.97568400239106, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 349, "compiled_prompt_sha256": "7b8ba837eec9ce754951db3caac4fd2da2ca873b5a41214f681f3059bf74da7a", "compiled_ids_sha256": "9a912b27d0a201b4913a165776dfd88073638d05b912990a61ae5735aa3462d6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.75, 18.25, 16.75, 18.875, 17.125, 19.25], "runtime_probabilities": [0.2118353396654129, 0.12848463654518127, 0.02866879478096962, 0.2400408834218979, 0.04171285033226013, 0.3492574095726013], "decide_seconds": 0.18270907799887937, "probabilities_f64_minus_runtime_max_abs": 4.4179369884655983e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.11821704987445025, "darwin": 0.008044771204273563, "sagan": 0.01599893792815549, "ibn_sina": 0.6802921515570143, "einstein": 0.13395746719135715, "socrates": 0.043489622244749275}, "confidence": 0.42788927412989064, "choice": "ibn_sina"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 186.55132800267893, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 349, "compiled_prompt_sha256": "4f5ec06b6ec67c809734ca47b5b836ce712df5bf1d4eed101c7b9f244a430910", "compiled_ids_sha256": "00d68204ee82e667e49ffce720b827a5438f82b83114a757fb965a9142f3e1de", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 15.4375, 16.125, 19.875, 18.25, 17.125], "runtime_probabilities": [0.11821704357862473, 0.008044770918786526, 0.01599893905222416, 0.6802921891212463, 0.13395746052265167, 0.04348962381482124], "decide_seconds": 0.18241332999969018, "probabilities_f64_minus_runtime_max_abs": 3.7564232058251434e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.23152413124457064, "einstein": 0.05166001648282714, "darwin": 0.024402463900537054, "sagan": 0.027651614219912635, "hypatia": 0.5553975183997963, "socrates": 0.1093642557523561}, "confidence": 0.30220724061710713, "choice": "hypatia"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 186.55989799663075, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 349, "compiled_prompt_sha256": "dc1f71e89449ba9f6ec5ec75538f0ac800608cc1bf7eabab422b36013ac66370", "compiled_ids_sha256": "f2547ae4afc2cc6b367dd1a27eba35253a4b8f94069ad1981f0f2976a1e99fd8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.75, 17.25, 16.5, 16.625, 19.625, 18.0], "runtime_probabilities": [0.23152412474155426, 0.05166001617908478, 0.02440246380865574, 0.02765161357820034, 0.5553975105285645, 0.10936424881219864], "decide_seconds": 0.18240866899941466, "probabilities_f64_minus_runtime_max_abs": 7.871231888501029e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.200031627180323, "sagan": 0.01641959584183305, "socrates": 0.15578478788708286, "ibn_sina": 0.5437423372813665, "darwin": 0.03938856280189819, "hypatia": 0.044633089007496476}, "confidence": 0.2875780994510855, "choice": "ibn_sina"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 186.49990799895022, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 349, "compiled_prompt_sha256": "56e7f69d049c3c61a0c93373197985ba905699c9fdca194d77cb1dc097f03228", "compiled_ids_sha256": "c54d8a0a1ed8ffd2b1ef1fd264e0e0bcbb3f064db7206763949f88f42737c4b9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.75, 16.25, 18.5, 19.75, 17.125, 17.25], "runtime_probabilities": [0.20003163814544678, 0.01641959510743618, 0.15578478574752808, 0.543742299079895, 0.039388563483953476, 0.044633086770772934], "decide_seconds": 0.18234656999993604, "probabilities_f64_minus_runtime_max_abs": 3.820147143063224e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.08430401554802439, "ibn_sina": 0.622927100249474, "hypatia": 0.0743980325965751, "einstein": 0.1784716023156828, "sagan": 0.03514314218835782, "darwin": 0.004756107101885958}, "confidence": 0.3596564103813238, "choice": "ibn_sina"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 186.42490899947006, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 349, "compiled_prompt_sha256": "5af5f17bf2c762b293eb16ba8999f89bbfdbe8db14a42dd8732b5bafceca03b6", "compiled_ids_sha256": "f906c10057f7a5b490afce1284e09e974c3469d23ce1191b29e087bb35b9907e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 19.75, 17.625, 18.5, 16.875, 14.875], "runtime_probabilities": [0.08430401235818863, 0.6229271292686462, 0.07439803332090378, 0.1784716099500656, 0.03514314442873001, 0.004756106995046139], "decide_seconds": 0.18228371099758078, "probabilities_f64_minus_runtime_max_abs": 2.9019172287370054e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:24Z"} {"id": "c-hypatia-01-main-03", "uid": "c046", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "prompt_sha256_by_order": ["51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe"], "request_sha256": "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "compiled_prompt_sha256_by_order": ["981f394b92cdbfe83d1629e86ff1043844d3ddec1aa960eee15412baa3b51c00", "214a087a51941e97f1c28690b67a09812e8d200430880e5f296a79504f2dfc50", "da154d0f83a59d3e823b138fe84d86d01b16328c89c63f73015febd9cabf2e5e", "81999bb5b2c7b6db7853aa583b6e07a90b67369699ec5fb2dc1fdd51e0c21d22", "532d14b926253ef4b4304a2a1f635221b18243ab97d9c594080fc80b1dcb97fc", "d725197c1e2a979cdabd3217a3603d86ad6a186762c6e489efbb13604dcf97d6"], "compiled_tokens_by_order": [364, 364, 364, 364, 364, 364], "choice": "ibn_sina", "probs": {"darwin": 0.08516241680920204, "einstein": 0.12391058530552576, "hypatia": 0.15910434092894968, "ibn_sina": 0.4324904387761166, "sagan": 0.15910434092894968, "socrates": 0.04022787725125622}, "confidence": 0.13759512058121448, "seconds": 0.1935662513345354, "seconds_all_orders": 1.1613975080072123, "seconds_by_order": [0.19366933200217318, 0.19358469300277648, 0.19356375300048967, 0.1937241220002761, 0.19337935400108108, 0.1934762540004158], "argmax_stable": false, "spread": {"darwin": 0.25716276813817945, "einstein": 0.26274640419379236, "hypatia": 0.1945223913127853, "ibn_sina": 0.2918511473561711, "sagan": 0.1208724319246673, "socrates": 0.030599984447701206}, "spread_2dp": {"darwin": 0.24999999999999997, "einstein": 0.26, "hypatia": 0.19, "ibn_sina": 0.29000000000000004, "sagan": 0.12, "socrates": 0.030000000000000002}, "choices_by_order": ["ibn_sina", "darwin", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.08516241680920204, "einstein": 0.12391058530552576, "hypatia": 0.15910434092894968, "ibn_sina": 0.4324904387761166, "sagan": 0.15910434092894968, "socrates": 0.04022787725125622}, {"ibn_sina": 0.22915339053245418, "hypatia": 0.1784648399901428, "sagan": 0.09552534519353684, "socrates": 0.04512297800846646, "darwin": 0.2942387777638432, "einstein": 0.15749466851155644}, {"hypatia": 0.11420214953961667, "darwin": 0.03707600962566378, "sagan": 0.06112800570251457, "ibn_sina": 0.4516782633516328, "einstein": 0.31043362786450257, "socrates": 0.025481943916069602}, {"ibn_sina": 0.35236357090513126, "einstein": 0.0476872236707102, "darwin": 0.18860662832073308, "sagan": 0.11439570270154968, "hypatia": 0.2744210249467527, "socrates": 0.022525849455123093}, {"einstein": 0.11625182597714763, "sagan": 0.042766656775632836, "socrates": 0.022891341798706586, "ibn_sina": 0.5210045378886253, "darwin": 0.21718700392592027, "hypatia": 0.0798986336339674}, {"socrates": 0.0531258339028243, "ibn_sina": 0.3464239534970184, "hypatia": 0.2101167490548052, "einstein": 0.09925212442416718, "sagan": 0.16363908870030014, "darwin": 0.12744225042088486}], "prompt_tokens": 364, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.08516241680920204, "einstein": 0.12391058530552576, "hypatia": 0.15910434092894968, "ibn_sina": 0.4324904387761166, "sagan": 0.15910434092894968, "socrates": 0.04022787725125622}, "confidence": 0.13759512058121448, "choice": "ibn_sina"}}, "usage": {"input_tokens": 364, "output_tokens": 1}, "latency_ms": 193.64252199738985, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 364, "compiled_prompt_sha256": "981f394b92cdbfe83d1629e86ff1043844d3ddec1aa960eee15412baa3b51c00", "compiled_ids_sha256": "e1be0ad9a565296219335b76cdbbdeef9cf60282395d86cca26940dc6907053b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 18.0, 18.25, 19.25, 18.25, 16.875], "runtime_probabilities": [0.08516241610050201, 0.12391059100627899, 0.1591043621301651, 0.43249043822288513, 0.1591043621301651, 0.040227875113487244], "decide_seconds": 0.18927872700078296, "probabilities_f64_minus_runtime_max_abs": 2.120121542215081e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.22915339053245418, "hypatia": 0.1784648399901428, "sagan": 0.09552534519353684, "socrates": 0.04512297800846646, "darwin": 0.2942387777638432, "einstein": 0.15749466851155644}, "confidence": 0.07331858261001256, "choice": "darwin"}}, "usage": {"input_tokens": 364, "output_tokens": 1}, "latency_ms": 193.56136300120852, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 364, "compiled_prompt_sha256": "214a087a51941e97f1c28690b67a09812e8d200430880e5f296a79504f2dfc50", "compiled_ids_sha256": "b5d6e2a37517c392d89a5125488f28128d5b96ea05775c430855018536e03c59", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.5, 18.25, 17.625, 16.875, 18.75, 18.125], "runtime_probabilities": [0.22915340960025787, 0.1784648597240448, 0.0955253466963768, 0.04512298107147217, 0.294238805770874, 0.1574946790933609], "decide_seconds": 0.18936526499965112, "probabilities_f64_minus_runtime_max_abs": 2.8007030805632382e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.11420214953961667, "darwin": 0.03707600962566378, "sagan": 0.06112800570251457, "ibn_sina": 0.4516782633516328, "einstein": 0.31043362786450257, "socrates": 0.025481943916069602}, "confidence": 0.24296069146645816, "choice": "ibn_sina"}}, "usage": {"input_tokens": 364, "output_tokens": 1}, "latency_ms": 193.54042399936588, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 364, "compiled_prompt_sha256": "da154d0f83a59d3e823b138fe84d86d01b16328c89c63f73015febd9cabf2e5e", "compiled_ids_sha256": "2cb0828d9884ce5b5b4b7bb65767edfb2d72c7c8f63be63b02de221c96abfddd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.875, 16.75, 17.25, 19.25, 18.875, 16.375], "runtime_probabilities": [0.11420214176177979, 0.03707600757479668, 0.061128001660108566, 0.45167824625968933, 0.31043362617492676, 0.025481944903731346], "decide_seconds": 0.1893547360014054, "probabilities_f64_minus_runtime_max_abs": 1.7091943471658766e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.35236357090513126, "einstein": 0.0476872236707102, "darwin": 0.18860662832073308, "sagan": 0.11439570270154968, "hypatia": 0.2744210249467527, "socrates": 0.022525849455123093}, "confidence": 0.15413207610283686, "choice": "ibn_sina"}}, "usage": {"input_tokens": 364, "output_tokens": 1}, "latency_ms": 193.70041199726984, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 364, "compiled_prompt_sha256": "81999bb5b2c7b6db7853aa583b6e07a90b67369699ec5fb2dc1fdd51e0c21d22", "compiled_ids_sha256": "97ed5bb9478331d4df11428579ccac8270e79aab1730f49c36a0bdd0086fe500", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.875, 16.875, 18.25, 17.75, 18.625, 16.125], "runtime_probabilities": [0.35236358642578125, 0.04768722504377365, 0.1886066496372223, 0.11439570784568787, 0.2744210362434387, 0.0225258506834507], "decide_seconds": 0.18946505500207422, "probabilities_f64_minus_runtime_max_abs": 2.1316489212663825e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.11625182597714763, "sagan": 0.042766656775632836, "socrates": 0.022891341798706586, "ibn_sina": 0.5210045378886253, "darwin": 0.21718700392592027, "hypatia": 0.0798986336339674}, "confidence": 0.24952263910083539, "choice": "ibn_sina"}}, "usage": {"input_tokens": 364, "output_tokens": 1}, "latency_ms": 193.35592400238966, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 364, "compiled_prompt_sha256": "532d14b926253ef4b4304a2a1f635221b18243ab97d9c594080fc80b1dcb97fc", "compiled_ids_sha256": "3ce4781acdffa8bac9ded1b451485eb8cd5da890f2f1df0aaa4feb47d745f7b6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 16.75, 16.125, 19.25, 18.375, 17.375], "runtime_probabilities": [0.11625182628631592, 0.04276665672659874, 0.022891342639923096, 0.5210045576095581, 0.21718700230121613, 0.07989863306283951], "decide_seconds": 0.18916655700013507, "probabilities_f64_minus_runtime_max_abs": 1.972093277569087e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0531258339028243, "ibn_sina": 0.3464239534970184, "hypatia": 0.2101167490548052, "einstein": 0.09925212442416718, "sagan": 0.16363908870030014, "darwin": 0.12744225042088486}, "confidence": 0.08525780464045163, "choice": "ibn_sina"}}, "usage": {"input_tokens": 364, "output_tokens": 1}, "latency_ms": 193.45150400113198, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 364, "compiled_prompt_sha256": "d725197c1e2a979cdabd3217a3603d86ad6a186762c6e489efbb13604dcf97d6", "compiled_ids_sha256": "f1dc423f6d5a6e2d18d92dc8aa688fe9ac56e2c0166b826629d31f4318efa327", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 18.875, 18.375, 17.625, 18.125, 17.875], "runtime_probabilities": [0.05312582850456238, 0.3464239239692688, 0.21011672914028168, 0.0992521196603775, 0.16363908350467682, 0.12744225561618805], "decide_seconds": 0.1892711260006763, "probabilities_f64_minus_runtime_max_abs": 2.9527749578761586e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:25Z"} {"id": "c-hypatia-01-main-03", "uid": "c047", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "prompt_sha256_by_order": ["ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f"], "request_sha256": "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "compiled_prompt_sha256_by_order": ["5e16fbf7126da6744631cdedb3c9e498a2b9052c185376919d00b1f279b12f9c", "8a9783359ca22b2ca060f289690fe6ed73f55643bd619baf4637a1b065c7ecca", "9878e42d49f4c5957122bb2b720a9531598548b6d8d8120ede38748a5a1a24de", "eb5b9290878f211609814ff1f2be0464ce38ec64a9a7a6beac89e8e9aa525329", "4440f5d875c123e4ff6fb95694a3ffa472d0c6be75bddc484d62adea6dfb75bb", "977d0b133c97885600b1c793fd8b912d7fdad5da8c8e2667e501ef1416aeaa88"], "compiled_tokens_by_order": [329, 329, 329, 329, 329, 329], "choice": "ibn_sina", "probs": {"darwin": 0.011407948343604452, "einstein": 0.029131216628354755, "hypatia": 0.04802925649654176, "ibn_sina": 0.8513389416396695, "sagan": 0.022687414321985414, "socrates": 0.037405222569844084}, "confidence": 0.6396408682512509, "seconds": 0.1848494389996631, "seconds_all_orders": 1.1090966339979786, "seconds_by_order": [0.18520826799795032, 0.1844764839988784, 0.18491789999825414, 0.18465922200266505, 0.1850107089994708, 0.18482405100075994], "argmax_stable": true, "spread": {"darwin": 0.01831428961549014, "einstein": 0.06054490964110412, "hypatia": 0.11290693997383447, "ibn_sina": 0.16359488594077298, "sagan": 0.025864671814127133, "socrates": 0.08041733398904606}, "spread_2dp": {"darwin": 0.02, "einstein": 0.06, "hypatia": 0.10999999999999999, "ibn_sina": 0.16000000000000003, "sagan": 0.03, "socrates": 0.08}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.011407948343604452, "einstein": 0.029131216628354755, "hypatia": 0.04802925649654176, "ibn_sina": 0.8513389416396695, "sagan": 0.022687414321985414, "socrates": 0.037405222569844084}, {"ibn_sina": 0.8115608002606082, "hypatia": 0.05878926718940113, "sagan": 0.040405233047253276, "socrates": 0.03565749301240975, "darwin": 0.017929713477917792, "einstein": 0.03565749301240975}, {"hypatia": 0.11437037328364055, "darwin": 0.007311453349435909, "sagan": 0.014540561233126142, "ibn_sina": 0.7457885169072556, "einstein": 0.08907173627346815, "socrates": 0.028917358953073652}, {"ibn_sina": 0.833676076882297, "einstein": 0.028526826632364024, "darwin": 0.015269309974352743, "sagan": 0.032325129469368837, "hypatia": 0.07754395445835667, "socrates": 0.012658702583260847}, {"einstein": 0.052066590322974976, "sagan": 0.019154228151718463, "socrates": 0.052066590322974976, "ibn_sina": 0.8144585058871318, "darwin": 0.021704583999808832, "hypatia": 0.040549501315390944}, {"socrates": 0.09307603657230691, "ibn_sina": 0.6877440556988965, "hypatia": 0.15345644128922542, "einstein": 0.038799864589202565, "sagan": 0.023533307466049876, "darwin": 0.003390294384318695}], "prompt_tokens": 329, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.011407948343604452, "einstein": 0.029131216628354755, "hypatia": 0.04802925649654176, "ibn_sina": 0.8513389416396695, "sagan": 0.022687414321985414, "socrates": 0.037405222569844084}, "confidence": 0.6396408682512509, "choice": "ibn_sina"}}, "usage": {"input_tokens": 329, "output_tokens": 1}, "latency_ms": 185.17744799828506, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 329, "compiled_prompt_sha256": "5e16fbf7126da6744631cdedb3c9e498a2b9052c185376919d00b1f279b12f9c", "compiled_ids_sha256": "07c699060a078fa6ae02cca28988321df21cdd9ed54996d09648ee655e350dce", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.4375, 16.375, 16.875, 19.75, 16.125, 16.625], "runtime_probabilities": [0.011407949030399323, 0.029131216928362846, 0.04802925884723663, 0.8513389825820923, 0.022687414661049843, 0.03740522265434265], "decide_seconds": 0.18093735200091032, "probabilities_f64_minus_runtime_max_abs": 4.094242278540605e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.8115608002606082, "hypatia": 0.05878926718940113, "sagan": 0.040405233047253276, "socrates": 0.03565749301240975, "darwin": 0.017929713477917792, "einstein": 0.03565749301240975}, "confidence": 0.567157381151714, "choice": "ibn_sina"}}, "usage": {"input_tokens": 329, "output_tokens": 1}, "latency_ms": 184.45060399972135, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 329, "compiled_prompt_sha256": "8a9783359ca22b2ca060f289690fe6ed73f55643bd619baf4637a1b065c7ecca", "compiled_ids_sha256": "2c872e1638c278cbcd624b5ddc049e1eb933fbc9162b13b45a35427b4454fd40", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.5, 16.875, 16.5, 16.375, 15.6875, 16.375], "runtime_probabilities": [0.811560869216919, 0.05878927558660507, 0.040405236184597015, 0.03565749526023865, 0.017929714173078537, 0.03565749526023865], "decide_seconds": 0.18046465500083286, "probabilities_f64_minus_runtime_max_abs": 6.895631077252773e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.11437037328364055, "darwin": 0.007311453349435909, "sagan": 0.014540561233126142, "ibn_sina": 0.7457885169072556, "einstein": 0.08907173627346815, "socrates": 0.028917358953073652}, "confidence": 0.5076987390324249, "choice": "ibn_sina"}}, "usage": {"input_tokens": 329, "output_tokens": 1}, "latency_ms": 184.89319100262946, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 329, "compiled_prompt_sha256": "9878e42d49f4c5957122bb2b720a9531598548b6d8d8120ede38748a5a1a24de", "compiled_ids_sha256": "edd99e6f4b88a96477c8055720e7d0ec32a51228358aced1e1cb659383548413", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.875, 15.125, 15.8125, 19.75, 17.625, 16.5], "runtime_probabilities": [0.11437037587165833, 0.007311453577131033, 0.014540561474859715, 0.7457885146141052, 0.08907173573970795, 0.028917357325553894], "decide_seconds": 0.1809738509982708, "probabilities_f64_minus_runtime_max_abs": 2.5880177723092146e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.833676076882297, "einstein": 0.028526826632364024, "darwin": 0.015269309974352743, "sagan": 0.032325129469368837, "hypatia": 0.07754395445835667, "socrates": 0.012658702583260847}, "confidence": 0.619648998506828, "choice": "ibn_sina"}}, "usage": {"input_tokens": 329, "output_tokens": 1}, "latency_ms": 184.63414299912984, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 329, "compiled_prompt_sha256": "eb5b9290878f211609814ff1f2be0464ce38ec64a9a7a6beac89e8e9aa525329", "compiled_ids_sha256": "681e7bdff5abe93a261513b6c3c8b6333c4c25ec3dfaac932bec6572fc7d713b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.625, 16.25, 15.625, 16.375, 17.25, 15.4375], "runtime_probabilities": [0.8336760997772217, 0.028526825830340385, 0.015269310213625431, 0.032325129956007004, 0.0775439515709877, 0.012658702209591866], "decide_seconds": 0.1805778050002118, "probabilities_f64_minus_runtime_max_abs": 2.2894924733840583e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.052066590322974976, "sagan": 0.019154228151718463, "socrates": 0.052066590322974976, "ibn_sina": 0.8144585058871318, "darwin": 0.021704583999808832, "hypatia": 0.040549501315390944}, "confidence": 0.5737407654758161, "choice": "ibn_sina"}}, "usage": {"input_tokens": 329, "output_tokens": 1}, "latency_ms": 184.9845689976064, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 329, "compiled_prompt_sha256": "4440f5d875c123e4ff6fb95694a3ffa472d0c6be75bddc484d62adea6dfb75bb", "compiled_ids_sha256": "2cf7750a23fd10fcd6564f42db1e9309b8cbe4df9456aef8d74e608547e1e0aa", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 16.0, 17.0, 19.75, 16.125, 16.75], "runtime_probabilities": [0.05206659063696861, 0.019154230132699013, 0.05206659063696861, 0.8144585490226746, 0.02170458436012268, 0.04054950177669525], "decide_seconds": 0.1808923419994244, "probabilities_f64_minus_runtime_max_abs": 4.3135542804328963e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.09307603657230691, "ibn_sina": 0.6877440556988965, "hypatia": 0.15345644128922542, "einstein": 0.038799864589202565, "sagan": 0.023533307466049876, "darwin": 0.003390294384318695}, "confidence": 0.4420787293811761, "choice": "ibn_sina"}}, "usage": {"input_tokens": 329, "output_tokens": 1}, "latency_ms": 184.79829100033385, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 329, "compiled_prompt_sha256": "977d0b133c97885600b1c793fd8b912d7fdad5da8c8e2667e501ef1416aeaa88", "compiled_ids_sha256": "5135af406a15d382926a370f50fa4a3b44465ddb59226ca7a7f46aa4af9d3568", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 19.75, 18.25, 16.875, 16.375, 14.4375], "runtime_probabilities": [0.09307603538036346, 0.6877440214157104, 0.153456449508667, 0.038799867033958435, 0.023533307015895844, 0.0033902942668646574], "decide_seconds": 0.18074675299794762, "probabilities_f64_minus_runtime_max_abs": 3.428318606779612e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:26Z"} {"id": "c-hypatia-01-main-03", "uid": "c048", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "prompt_sha256_by_order": ["518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec"], "request_sha256": "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "compiled_prompt_sha256_by_order": ["1758a9f1601e86873bbd3a7f9387586fd8779e1252ba78aa05c488753ae18af9", "6a674a8716b378f53580b41381308be25f2606421b204bd7b8ffcc49badc30c9", "fc7fa9a6c7732de9f42f45ac6c843b49cf098778d44da80166146a77f93011ac", "1b54e6c415f7d855064ed9713ec139f4b68aca281ae2acdec906924e108ca2da", "0268fe9666c5c235d576b3d8accff25eab943327704299b9aeaa2e91014df2ec", "3e2a04c18bb16bfe4025d0adaece9bd3dc34c5b34f81f8b1db278206a2cedebe"], "compiled_tokens_by_order": [407, 407, 407, 407, 407, 407], "choice": "hypatia", "probs": {"darwin": 0.0027728356856710676, "einstein": 0.0008456671045967907, "hypatia": 0.987197884086869, "ibn_sina": 0.0051803390603787774, "sagan": 0.0027728356856710676, "socrates": 0.0012304383768133628}, "confidence": 0.9515205859838446, "seconds": 0.2132313946664605, "seconds_all_orders": 1.279388367998763, "seconds_by_order": [0.21340342899929965, 0.2132951290004712, 0.21282460300062667, 0.21298894099891186, 0.2134383179982251, 0.2134379480012285], "argmax_stable": true, "spread": {"darwin": 0.008198354700540863, "einstein": 0.004563934016839152, "hypatia": 0.018768700707517016, "ibn_sina": 0.017977697722621686, "sagan": 0.0017547218681712646, "socrates": 0.0013472010615958655}, "spread_2dp": {"darwin": 0.01, "einstein": 0.01, "hypatia": 0.020000000000000018, "ibn_sina": 0.02, "sagan": 0.0, "socrates": 0.0}, "choices_by_order": ["hypatia", "hypatia", "hypatia", "hypatia", "hypatia", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0027728356856710676, "einstein": 0.0008456671045967907, "hypatia": 0.987197884086869, "ibn_sina": 0.0051803390603787774, "sagan": 0.0027728356856710676, "socrates": 0.0012304383768133628}, {"ibn_sina": 0.006221993910169959, "hypatia": 0.9829816176493456, "sagan": 0.0024365678243713335, "socrates": 0.0022889436426454415, "darwin": 0.004845693729505876, "einstein": 0.00122518324396186}, {"hypatia": 0.9701752155035464, "darwin": 0.002559921745153182, "sagan": 0.0029007713654924587, "ibn_sina": 0.020135344219250062, "einstein": 0.003287004585508325, "socrates": 0.000941742581049576}, {"ibn_sina": 0.010814524984941997, "einstein": 0.0016584611205094305, "darwin": 0.008965556111806373, "sagan": 0.003510962219669939, "hypatia": 0.9734925155222974, "socrates": 0.0015579800407748332}, {"einstein": 0.005409601121435943, "sagan": 0.0017562403514986746, "socrates": 0.0014559743899157825, "ibn_sina": 0.012190724312103482, "darwin": 0.010758276445694045, "hypatia": 0.968429183379352}, {"socrates": 0.0021576464966283782, "ibn_sina": 0.0021576464966283782, "hypatia": 0.9863566890368458, "einstein": 0.0021576464966283782, "sagan": 0.00313935712837533, "darwin": 0.004031014344893764}], "prompt_tokens": 407, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0027728356856710676, "einstein": 0.0008456671045967907, "hypatia": 0.987197884086869, "ibn_sina": 0.0051803390603787774, "sagan": 0.0027728356856710676, "socrates": 0.0012304383768133628}, "confidence": 0.9515205859838446, "choice": "hypatia"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 213.37162900090334, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "1758a9f1601e86873bbd3a7f9387586fd8779e1252ba78aa05c488753ae18af9", "compiled_ids_sha256": "c2e4ce3277e1b394388c8fbe6816aea6f422bb1f7b5f4acc516099b73569fad1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.125, 13.9375, 21.0, 15.75, 15.125, 14.3125], "runtime_probabilities": [0.0027728357817977667, 0.0008456670329906046, 0.9871978163719177, 0.005180338863283396, 0.0027728357817977667, 0.0012304383562877774], "decide_seconds": 0.2083573579984659, "probabilities_f64_minus_runtime_max_abs": 6.771495131285832e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.006221993910169959, "hypatia": 0.9829816176493456, "sagan": 0.0024365678243713335, "socrates": 0.0022889436426454415, "darwin": 0.004845693729505876, "einstein": 0.00122518324396186}, "confidence": 0.937996146243518, "choice": "hypatia"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 213.26785899873357, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "6a674a8716b378f53580b41381308be25f2606421b204bd7b8ffcc49badc30c9", "compiled_ids_sha256": "1833fa669c887af0fa128aaaf50de0b1c67feaff6b6de7bac2408f82e6b84ee0", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 21.0, 15.0, 14.9375, 15.6875, 14.3125], "runtime_probabilities": [0.006221993826329708, 0.9829816222190857, 0.0024365680292248726, 0.0022889436222612858, 0.0048456937074661255, 0.0012251833686605096], "decide_seconds": 0.20852648700019927, "probabilities_f64_minus_runtime_max_abs": 4.569740141846523e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.9701752155035464, "darwin": 0.002559921745153182, "sagan": 0.0029007713654924587, "ibn_sina": 0.020135344219250062, "einstein": 0.003287004585508325, "socrates": 0.000941742581049576}, "confidence": 0.9075816271040877, "choice": "hypatia"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 212.79617300024256, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "fc7fa9a6c7732de9f42f45ac6c843b49cf098778d44da80166146a77f93011ac", "compiled_ids_sha256": "ba628d8ae8ec01a5e86b8ee55896704e2e2f7edda6088438ce8ad7aef2bc8ecc", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.875, 14.9375, 15.0625, 17.0, 15.1875, 13.9375], "runtime_probabilities": [0.9701752662658691, 0.0025599219370633364, 0.002900771563872695, 0.020135344937443733, 0.0032870047725737095, 0.0009417426190339029], "decide_seconds": 0.20809074000135297, "probabilities_f64_minus_runtime_max_abs": 5.076232278966586e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.010814524984941997, "einstein": 0.0016584611205094305, "darwin": 0.008965556111806373, "sagan": 0.003510962219669939, "hypatia": 0.9734925155222974, "socrates": 0.0015579800407748332}, "confidence": 0.9118699242423483, "choice": "hypatia"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 212.96261200041045, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "1b54e6c415f7d855064ed9713ec139f4b68aca281ae2acdec906924e108ca2da", "compiled_ids_sha256": "6cca4592ede3a8d5234c949aecc06cece05232538cfb1cc1a5f5adcce32c90ab", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 14.25, 15.9375, 15.0, 20.625, 14.1875], "runtime_probabilities": [0.010814526118338108, 0.0016584614058956504, 0.008965556509792805, 0.003510962473228574, 0.9734925627708435, 0.0015579801984131336], "decide_seconds": 0.20828304899987415, "probabilities_f64_minus_runtime_max_abs": 4.7248546075628894e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.005409601121435943, "sagan": 0.0017562403514986746, "socrates": 0.0014559743899157825, "ibn_sina": 0.012190724312103482, "darwin": 0.010758276445694045, "hypatia": 0.968429183379352}, "confidence": 0.8981788620175083, "choice": "hypatia"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 213.41052800198668, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "0268fe9666c5c235d576b3d8accff25eab943327704299b9aeaa2e91014df2ec", "compiled_ids_sha256": "dc81e9218eab3129e8be26b6f35a94acc39ea2c268947a6304cdda3dbc9f04d6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.8125, 14.6875, 14.5, 16.625, 16.5, 21.0], "runtime_probabilities": [0.005409601144492626, 0.0017562402645125985, 0.0014559742994606495, 0.012190723791718483, 0.010758277028799057, 0.9684291481971741], "decide_seconds": 0.2087463950010715, "probabilities_f64_minus_runtime_max_abs": 3.5182177948733795e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0021576464966283782, "ibn_sina": 0.0021576464966283782, "hypatia": 0.9863566890368458, "einstein": 0.0021576464966283782, "sagan": 0.00313935712837533, "darwin": 0.004031014344893764}, "confidence": 0.9477575675877754, "choice": "hypatia"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 213.41035799923702, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 407, "compiled_prompt_sha256": "3e2a04c18bb16bfe4025d0adaece9bd3dc34c5b34f81f8b1db278206a2cedebe", "compiled_ids_sha256": "c3d418fa8f304a9c23bf22affdb273fe008afd503efd7edd5ba1462d0fe2b0e7", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0, 15.0, 21.125, 15.0, 15.375, 15.625], "runtime_probabilities": [0.0021576466970145702, 0.0021576466970145702, 0.9863567352294922, 0.0021576466970145702, 0.003139357315376401, 0.004031014628708363], "decide_seconds": 0.2084478470023896, "probabilities_f64_minus_runtime_max_abs": 4.619264637462095e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:27Z"} {"id": "c-hypatia-01-starter-01", "uid": "c049", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "prompt_sha256_by_order": ["2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e"], "request_sha256": "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "compiled_prompt_sha256_by_order": ["86edd11982c21ff733510b65fb8061cf99f203a5783f0d267f8b7922baef66d7", "9ca6df4d96fbece59b6a36b86ab04b90e3c8ba3e79e2b24eb8a990f20c891ef4", "41d3926c51b47a1f4a3e6fb517b1394348ef7a9ff3fe5a47cc6fe56ec59913b5", "090512e592eb25c1371a09a8a7c86aa8936430c30e71b7b28dcd17c102105fbc", "2c5d1055ff0406775e9fa226204eb617ecd53903ae204b80e077dd04042fe029", "6e859a1ee6e11bc8c3827dfa923f46015a885917a8f0c733a1a062066aca9189"], "compiled_tokens_by_order": [272, 272, 272, 272, 272, 272], "choice": "ibn_sina", "probs": {"darwin": 0.02198022230060404, "einstein": 0.024906854897994424, "hypatia": 0.0869334655857388, "ibn_sina": 0.8248017568166548, "sagan": 0.02198022230060404, "socrates": 0.019397478098403913}, "confidence": 0.6051458509357546, "seconds": 0.17237606616799894, "seconds_all_orders": 1.0342563970079937, "seconds_by_order": [0.17210522000095807, 0.17227839900078834, 0.1726210960005119, 0.17246702700140304, 0.1715325040022435, 0.1732521510020888], "argmax_stable": true, "spread": {"darwin": 0.016938527078872435, "einstein": 0.0657030964852499, "hypatia": 0.08888802268550601, "ibn_sina": 0.1803150179775347, "sagan": 0.01626723315926446, "socrates": 0.03399596247628483}, "spread_2dp": {"darwin": 0.01, "einstein": 0.060000000000000005, "hypatia": 0.09, "ibn_sina": 0.17999999999999994, "sagan": 0.01, "socrates": 0.04}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.02198022230060404, "einstein": 0.024906854897994424, "hypatia": 0.0869334655857388, "ibn_sina": 0.8248017568166548, "sagan": 0.02198022230060404, "socrates": 0.019397478098403913}, {"ibn_sina": 0.9093719286169354, "hypatia": 0.03526017615652692, "sagan": 0.01772993001631978, "socrates": 0.014698628246829798, "darwin": 0.010753746149457259, "einstein": 0.012185590813930832}, {"hypatia": 0.10882831410103658, "darwin": 0.010122610047765563, "sagan": 0.007883496631926291, "ibn_sina": 0.8041385180446047, "einstein": 0.0582515988690342, "socrates": 0.010775462305632689}, {"ibn_sina": 0.960777045111097, "einstein": 0.006891180316897392, "darwin": 0.005041695221731604, "sagan": 0.005712989141339577, "hypatia": 0.019940291415530564, "socrates": 0.0016367987934039674}, {"einstein": 0.04575337113704698, "sagan": 0.013108560303288528, "socrates": 0.0356327612696888, "ibn_sina": 0.8109979083907993, "darwin": 0.01907284269928977, "hypatia": 0.07543455619988666}, {"socrates": 0.02670594198221894, "ibn_sina": 0.7804620271335623, "hypatia": 0.09321289652002242, "einstein": 0.07259427680214729, "sagan": 0.0183547076043928, "darwin": 0.008670149957656316}], "prompt_tokens": 272, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.02198022230060404, "einstein": 0.024906854897994424, "hypatia": 0.0869334655857388, "ibn_sina": 0.8248017568166548, "sagan": 0.02198022230060404, "socrates": 0.019397478098403913}, "confidence": 0.6051458509357546, "choice": "ibn_sina"}}, "usage": {"input_tokens": 272, "output_tokens": 1}, "latency_ms": 172.07824100114522, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 272, "compiled_prompt_sha256": "86edd11982c21ff733510b65fb8061cf99f203a5783f0d267f8b7922baef66d7", "compiled_ids_sha256": "0632451188e0cd8274e38703ba69b1d81b731677e7550055093537363c0d1780", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 16.25, 17.5, 19.75, 16.125, 16.0], "runtime_probabilities": [0.021980224177241325, 0.024906856939196587, 0.08693347126245499, 0.8248018026351929, 0.021980224177241325, 0.01939748041331768], "decide_seconds": 0.16860864699992817, "probabilities_f64_minus_runtime_max_abs": 4.581853807472669e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.9093719286169354, "hypatia": 0.03526017615652692, "sagan": 0.01772993001631978, "socrates": 0.014698628246829798, "darwin": 0.010753746149457259, "einstein": 0.012185590813930832}, "confidence": 0.7542581825042435, "choice": "ibn_sina"}}, "usage": {"input_tokens": 272, "output_tokens": 1}, "latency_ms": 172.25638899981277, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 272, "compiled_prompt_sha256": "9ca6df4d96fbece59b6a36b86ab04b90e3c8ba3e79e2b24eb8a990f20c891ef4", "compiled_ids_sha256": "466504b359a0b9c3df1f7bf4cb4364e60600e74cc215d009d40ebeb38e0a21f3", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.875, 16.625, 15.9375, 15.75, 15.4375, 15.5625], "runtime_probabilities": [0.9093719720840454, 0.03526017814874649, 0.017729932442307472, 0.014698628336191177, 0.010753746144473553, 0.012185591273009777], "decide_seconds": 0.16888862500127289, "probabilities_f64_minus_runtime_max_abs": 4.346711002245485e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.10882831410103658, "darwin": 0.010122610047765563, "sagan": 0.007883496631926291, "ibn_sina": 0.8041385180446047, "einstein": 0.0582515988690342, "socrates": 0.010775462305632689}, "confidence": 0.600522504209827, "choice": "ibn_sina"}}, "usage": {"input_tokens": 272, "output_tokens": 1}, "latency_ms": 172.59975599881727, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 272, "compiled_prompt_sha256": "41d3926c51b47a1f4a3e6fb517b1394348ef7a9ff3fe5a47cc6fe56ec59913b5", "compiled_ids_sha256": "70b1a9174b9a981f2b65e1fa234015055ab28c6563e08586231b72dd78c54f11", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.875, 15.5, 15.25, 19.875, 17.25, 15.5625], "runtime_probabilities": [0.10882832109928131, 0.010122611187398434, 0.007883497513830662, 0.8041386008262634, 0.058251604437828064, 0.010775463655591011], "decide_seconds": 0.16919324199989205, "probabilities_f64_minus_runtime_max_abs": 8.278165875008625e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.960777045111097, "einstein": 0.006891180316897392, "darwin": 0.005041695221731604, "sagan": 0.005712989141339577, "hypatia": 0.019940291415530564, "socrates": 0.0016367987934039674}, "confidence": 0.8786169141052078, "choice": "ibn_sina"}}, "usage": {"input_tokens": 272, "output_tokens": 1}, "latency_ms": 172.44628700063913, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 272, "compiled_prompt_sha256": "090512e592eb25c1371a09a8a7c86aa8936430c30e71b7b28dcd17c102105fbc", "compiled_ids_sha256": "def3bd76ed0253733af5bf1344c531965c9d65e5944e97016387231d2cc5138b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.125, 15.1875, 14.875, 15.0, 16.25, 13.75], "runtime_probabilities": [0.9607770442962646, 0.006891179829835892, 0.005041695199906826, 0.005712989252060652, 0.019940290600061417, 0.0016367989592254162], "decide_seconds": 0.16908065299867303, "probabilities_f64_minus_runtime_max_abs": 8.154691470696473e-10}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.04575337113704698, "sagan": 0.013108560303288528, "socrates": 0.0356327612696888, "ibn_sina": 0.8109979083907993, "darwin": 0.01907284269928977, "hypatia": 0.07543455619988666}, "confidence": 0.5774342593089591, "choice": "ibn_sina"}}, "usage": {"input_tokens": 272, "output_tokens": 1}, "latency_ms": 171.51085400109878, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 272, "compiled_prompt_sha256": "2c5d1055ff0406775e9fa226204eb617ecd53903ae204b80e077dd04042fe029", "compiled_ids_sha256": "8d923a5efcfabbc03486d7d11fba9defcd41cde57b8767d84921dfb0ee5276ae", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 15.5, 16.5, 19.625, 15.875, 17.25], "runtime_probabilities": [0.04575337469577789, 0.013108559884130955, 0.03563276305794716, 0.8109979033470154, 0.019072843715548515, 0.07543455064296722], "decide_seconds": 0.1681294509980944, "probabilities_f64_minus_runtime_max_abs": 5.556919438487462e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.02670594198221894, "ibn_sina": 0.7804620271335623, "hypatia": 0.09321289652002242, "einstein": 0.07259427680214729, "sagan": 0.0183547076043928, "darwin": 0.008670149957656316}, "confidence": 0.5443939874569783, "choice": "ibn_sina"}}, "usage": {"input_tokens": 272, "output_tokens": 1}, "latency_ms": 173.23083099836367, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 272, "compiled_prompt_sha256": "6e859a1ee6e11bc8c3827dfa923f46015a885917a8f0c733a1a062066aca9189", "compiled_ids_sha256": "7a99cb6ba0a11ada1276eb69a6695f8482f07b8bf729e14789d4bb0ebd66785f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 19.75, 17.625, 17.375, 16.0, 15.25], "runtime_probabilities": [0.026705941185355186, 0.7804620265960693, 0.09321289509534836, 0.07259427756071091, 0.018354708328843117, 0.008670150302350521], "decide_seconds": 0.16989940700295847, "probabilities_f64_minus_runtime_max_abs": 1.424674059280484e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:28Z"} {"id": "c-hypatia-01-starter-01", "uid": "c050", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "prompt_sha256_by_order": ["49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7"], "request_sha256": "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "compiled_prompt_sha256_by_order": ["ecbc5708033f479cd2caa368a60bec322a499a7331fb8a6eacdc2dac8b96a819", "8f0580154b4096fd98b33a416ca21c762667e74d33ce54f82dd54523370fe337", "d2bb427dc7e60d790490212202c7e2fbe8521778e243040981565edde2647c70", "a3b372a62776ed9ecaba5285e18e91a63a3492e8a908a6b398380a222cfd691e", "1ff82cbd9a8fa813d552587570252c629dc21bc8e01a98dacff34dfa6fc901b8", "c87a85d26c8641910f485c0600d3d5a19d21cbf7c411c8c88c492ba178ddad9a"], "compiled_tokens_by_order": [326, 326, 326, 326, 326, 326], "choice": "ibn_sina", "probs": {"darwin": 0.05114647548749163, "einstein": 0.0073683484088708315, "hypatia": 0.0451366061958303, "ibn_sina": 0.8000654882244853, "sagan": 0.05114647548749163, "socrates": 0.0451366061958303}, "confidence": 0.5543810750994516, "seconds": 0.1844208626668357, "seconds_all_orders": 1.1065251760010142, "seconds_by_order": [0.18450196300182142, 0.1845123439998133, 0.18413381700156606, 0.18429359499714337, 0.18462619300044025, 0.18445726400022977], "argmax_stable": false, "spread": {"darwin": 0.2678598992361937, "einstein": 0.04996015629487646, "hypatia": 0.33284993183274214, "ibn_sina": 0.7518507222534081, "sagan": 0.0681076182877072, "socrates": 0.38293684109363046}, "spread_2dp": {"darwin": 0.26, "einstein": 0.049999999999999996, "hypatia": 0.33999999999999997, "ibn_sina": 0.76, "sagan": 0.060000000000000005, "socrates": 0.38}, "choices_by_order": ["ibn_sina", "socrates", "ibn_sina", "hypatia", "ibn_sina", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.05114647548749163, "einstein": 0.0073683484088708315, "hypatia": 0.0451366061958303, "ibn_sina": 0.8000654882244853, "sagan": 0.05114647548749163, "socrates": 0.0451366061958303}, {"ibn_sina": 0.22094753084652058, "hypatia": 0.033883401267436976, "sagan": 0.06330252744194298, "socrates": 0.3642808938153312, "darwin": 0.2837022453613312, "einstein": 0.033883401267436976}, {"hypatia": 0.10330507352368448, "darwin": 0.01584234612513753, "sagan": 0.01686408966646772, "ibn_sina": 0.76332698357066, "einstein": 0.038003812718067805, "socrates": 0.06265769439598246}, {"ibn_sina": 0.32364103053543275, "einstein": 0.01825857752783324, "darwin": 0.09272470770917673, "sagan": 0.06372869748754378, "hypatia": 0.36673333310017914, "socrates": 0.13491365363983432}, {"einstein": 0.014960441483442957, "sagan": 0.005503638852607059, "socrates": 0.04066669623016787, "ibn_sina": 0.8168124286750936, "darwin": 0.07597539083413533, "hypatia": 0.04608140392455327}, {"socrates": 0.42360353732379835, "ibn_sina": 0.06496170642168551, "hypatia": 0.3299027665795912, "einstein": 0.05732850470374729, "sagan": 0.07361125714031426, "darwin": 0.05059222783086339}], "prompt_tokens": 326, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.05114647548749163, "einstein": 0.0073683484088708315, "hypatia": 0.0451366061958303, "ibn_sina": 0.8000654882244853, "sagan": 0.05114647548749163, "socrates": 0.0451366061958303}, "confidence": 0.5543810750994516, "choice": "ibn_sina"}}, "usage": {"input_tokens": 326, "output_tokens": 1}, "latency_ms": 184.47640299928025, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 326, "compiled_prompt_sha256": "ecbc5708033f479cd2caa368a60bec322a499a7331fb8a6eacdc2dac8b96a819", "compiled_ids_sha256": "3213fbf2a47076131e60a1856d79082d20d04b59cb63f46819df660c2b4096ff", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 15.1875, 17.0, 19.875, 17.125, 17.0], "runtime_probabilities": [0.051146477460861206, 0.00736834853887558, 0.04513661190867424, 0.8000655174255371, 0.051146477460861206, 0.04513661190867424], "decide_seconds": 0.18045445499956259, "probabilities_f64_minus_runtime_max_abs": 2.920105179882171e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.22094753084652058, "hypatia": 0.033883401267436976, "sagan": 0.06330252744194298, "socrates": 0.3642808938153312, "darwin": 0.2837022453613312, "einstein": 0.033883401267436976}, "confidence": 0.18350930827008138, "choice": "socrates"}}, "usage": {"input_tokens": 326, "output_tokens": 1}, "latency_ms": 184.48884399913368, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 326, "compiled_prompt_sha256": "8f0580154b4096fd98b33a416ca21c762667e74d33ce54f82dd54523370fe337", "compiled_ids_sha256": "3c5d83ceabce936b6414b34a3cdca48c14d190eac44812367c72567d2a62cd9b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.25, 16.375, 17.0, 18.75, 18.5, 16.375], "runtime_probabilities": [0.2209475338459015, 0.03388340026140213, 0.06330252438783646, 0.36428090929985046, 0.2837022542953491, 0.03388340026140213], "decide_seconds": 0.18058449400268728, "probabilities_f64_minus_runtime_max_abs": 1.5484519277908504e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.10330507352368448, "darwin": 0.01584234612513753, "sagan": 0.01686408966646772, "ibn_sina": 0.76332698357066, "einstein": 0.038003812718067805, "socrates": 0.06265769439598246}, "confidence": 0.5127593525380324, "choice": "ibn_sina"}}, "usage": {"input_tokens": 326, "output_tokens": 1}, "latency_ms": 184.11127700164798, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 326, "compiled_prompt_sha256": "d2bb427dc7e60d790490212202c7e2fbe8521778e243040981565edde2647c70", "compiled_ids_sha256": "18e44558ab61bf0094ab3491ca6525c0e8ceff8031aac6cfedaf27a0b4fc2ba2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 15.875, 15.9375, 19.75, 16.75, 17.25], "runtime_probabilities": [0.10330507159233093, 0.015842346474528313, 0.01686408929526806, 0.7633270025253296, 0.03800381347537041, 0.06265769898891449], "decide_seconds": 0.1803327259985963, "probabilities_f64_minus_runtime_max_abs": 1.895466961165937e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.32364103053543275, "einstein": 0.01825857752783324, "darwin": 0.09272470770917673, "sagan": 0.06372869748754378, "hypatia": 0.36673333310017914, "socrates": 0.13491365363983432}, "confidence": 0.17830124976398898, "choice": "hypatia"}}, "usage": {"input_tokens": 326, "output_tokens": 1}, "latency_ms": 184.27044599957298, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 326, "compiled_prompt_sha256": "a3b372a62776ed9ecaba5285e18e91a63a3492e8a908a6b398380a222cfd691e", "compiled_ids_sha256": "fe1689c093e36ce7127e581f5f58900e0f5411ab3498d73285e79b5f1e8648cf", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.875, 16.0, 17.625, 17.25, 19.0, 18.0], "runtime_probabilities": [0.3236410319805145, 0.018258577212691307, 0.09272471070289612, 0.0637286975979805, 0.3667333424091339, 0.13491366803646088], "decide_seconds": 0.1804979850021482, "probabilities_f64_minus_runtime_max_abs": 1.4396626557555692e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.014960441483442957, "sagan": 0.005503638852607059, "socrates": 0.04066669623016787, "ibn_sina": 0.8168124286750936, "darwin": 0.07597539083413533, "hypatia": 0.04608140392455327}, "confidence": 0.5955754961576514, "choice": "ibn_sina"}}, "usage": {"input_tokens": 326, "output_tokens": 1}, "latency_ms": 184.60266300098738, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 326, "compiled_prompt_sha256": "1ff82cbd9a8fa813d552587570252c629dc21bc8e01a98dacff34dfa6fc901b8", "compiled_ids_sha256": "ef41420cca1207458ea2379a7b8678012efd36f35122fbf25370351e2cbee376", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 15.0, 17.0, 20.0, 17.625, 17.125], "runtime_probabilities": [0.01496044173836708, 0.0055036386474967, 0.040666695684194565, 0.8168123960494995, 0.07597538828849792, 0.04608140513300896], "decide_seconds": 0.18081752299985965, "probabilities_f64_minus_runtime_max_abs": 3.262559411520982e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.42360353732379835, "ibn_sina": 0.06496170642168551, "hypatia": 0.3299027665795912, "einstein": 0.05732850470374729, "sagan": 0.07361125714031426, "darwin": 0.05059222783086339}, "confidence": 0.21070800870081297, "choice": "socrates"}}, "usage": {"input_tokens": 326, "output_tokens": 1}, "latency_ms": 184.43473400111543, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 326, "compiled_prompt_sha256": "c87a85d26c8641910f485c0600d3d5a19d21cbf7c411c8c88c492ba178ddad9a", "compiled_ids_sha256": "ce26786a3b5b708763889c8e40308deb955fcfd36bbcd5adc0a035bc38d423c1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.25, 17.375, 19.0, 17.25, 17.5, 17.125], "runtime_probabilities": [0.42360353469848633, 0.06496170908212662, 0.3299027681350708, 0.05732850357890129, 0.07361125200986862, 0.05059222877025604], "decide_seconds": 0.18062730300152907, "probabilities_f64_minus_runtime_max_abs": 5.130445637768766e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:29Z"} {"id": "c-hypatia-01-starter-02", "uid": "c051", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "prompt_sha256_by_order": ["d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd"], "request_sha256": "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "compiled_prompt_sha256_by_order": ["d3cc18c537cd2524953ff31d3b343b481a694d8832a6516cf95f0af6ac3a93fc", "e5e94315a854703a398e48adb212aee85dbcddd9bfa474bab8d009899505f94f", "f5114197072d0456f7f044d5682ca0a5e97b2cc3e89232c198233ea48eea91a3", "316e4e233f9cec8ca26a21467e737810ce71f6e02d434c3980ecd0a7a99e5cdb", "2b8889e768b472422cf59693e3e301d0243bcba7fc27cbe5e6a0f9fbb85d85ce", "b0c5bf7d8ff8c34974293c92ef2fbd6781c4f2e077caba91ae21a56bb5a9ee7b"], "compiled_tokens_by_order": [303, 303, 303, 303, 303, 303], "choice": "hypatia", "probs": {"darwin": 0.0015888775617803328, "einstein": 0.0005845153895175926, "hypatia": 0.9326479803152403, "ibn_sina": 0.05962219064030882, "sagan": 0.004319017003833792, "socrates": 0.0012374190893191077}, "confidence": 0.8439852053774726, "seconds": 0.17549225716781316, "seconds_all_orders": 1.052953543006879, "seconds_by_order": [0.17536728500272147, 0.17521138600204722, 0.17549406399848522, 0.17564852199939196, 0.1752569360032794, 0.17597535000095377], "argmax_stable": true, "spread": {"darwin": 0.004383545085983665, "einstein": 0.0038961656532668695, "hypatia": 0.04817562577870271, "ibn_sina": 0.054420059658633536, "sagan": 0.010392952101177227, "socrates": 0.006210223623816556}, "spread_2dp": {"darwin": 0.01, "einstein": 0.0, "hypatia": 0.04999999999999993, "ibn_sina": 0.060000000000000005, "sagan": 0.01, "socrates": 0.01}, "choices_by_order": ["hypatia", "hypatia", "hypatia", "hypatia", "hypatia", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0015888775617803328, "einstein": 0.0005845153895175926, "hypatia": 0.9326479803152403, "ibn_sina": 0.05962219064030882, "sagan": 0.004319017003833792, "socrates": 0.0012374190893191077}, {"ibn_sina": 0.022401622000438104, "hypatia": 0.9525412060150392, "sagan": 0.01199072419309694, "socrates": 0.0064181721605742845, "darwin": 0.005664017051961487, "einstein": 0.0009842585788899124}, {"hypatia": 0.9650971148155055, "darwin": 0.0012804719659778223, "sagan": 0.002546522577489797, "ibn_sina": 0.029143407615857164, "einstein": 0.0012028920914060187, "socrates": 0.0007295909337636019}, {"ibn_sina": 0.06659981351635547, "einstein": 0.002582355010837096, "darwin": 0.0027489025999948573, "sagan": 0.007019568700588639, "hypatia": 0.9193820664647633, "socrates": 0.0016672937074606541}, {"einstein": 0.0026342808336790725, "sagan": 0.0015977720919197137, "socrates": 0.0018105129743121707, "ibn_sina": 0.0529109449509066, "darwin": 0.0031775492270536164, "hypatia": 0.9378689399221288}, {"socrates": 0.006939814557580157, "ibn_sina": 0.01217975385772193, "hypatia": 0.967557692243466, "einstein": 0.004480681042784462, "sagan": 0.007387394142461573, "darwin": 0.0014546641559857589}], "prompt_tokens": 303, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0015888775617803328, "einstein": 0.0005845153895175926, "hypatia": 0.9326479803152403, "ibn_sina": 0.05962219064030882, "sagan": 0.004319017003833792, "socrates": 0.0012374190893191077}, "confidence": 0.8439852053774726, "choice": "hypatia"}}, "usage": {"input_tokens": 303, "output_tokens": 1}, "latency_ms": 175.34109500047634, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 303, "compiled_prompt_sha256": "d3cc18c537cd2524953ff31d3b343b481a694d8832a6516cf95f0af6ac3a93fc", "compiled_ids_sha256": "4176637f3b7f7e0c1eaf039089f3e571cafaabcf2dff27018fe9ae1943bf4ef4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.0, 13.0, 20.375, 17.625, 15.0, 13.75], "runtime_probabilities": [0.0015888777561485767, 0.0005845153355039656, 0.9326480031013489, 0.05962218716740608, 0.0043190172873437405, 0.0012374190846458077], "decide_seconds": 0.1715170540010149, "probabilities_f64_minus_runtime_max_abs": 2.27861085555503e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.022401622000438104, "hypatia": 0.9525412060150392, "sagan": 0.01199072419309694, "socrates": 0.0064181721605742845, "darwin": 0.005664017051961487, "einstein": 0.0009842585788899124}, "confidence": 0.8588130030152417, "choice": "hypatia"}}, "usage": {"input_tokens": 303, "output_tokens": 1}, "latency_ms": 175.189017001685, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 303, "compiled_prompt_sha256": "e5e94315a854703a398e48adb212aee85dbcddd9bfa474bab8d009899505f94f", "compiled_ids_sha256": "661f53f3bee1418a76f3203da4a844064802ab82de09e99b9e4e14f4acf44db4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 20.5, 16.125, 15.5, 15.375, 13.625], "runtime_probabilities": [0.02240162156522274, 0.952541172504425, 0.011990724131464958, 0.006418171804398298, 0.0056640165857970715, 0.000984258484095335], "decide_seconds": 0.17150854500141577, "probabilities_f64_minus_runtime_max_abs": 3.351061417244949e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.9650971148155055, "darwin": 0.0012804719659778223, "sagan": 0.002546522577489797, "ibn_sina": 0.029143407615857164, "einstein": 0.0012028920914060187, "socrates": 0.0007295909337636019}, "confidence": 0.9026544268249297, "choice": "hypatia"}}, "usage": {"input_tokens": 303, "output_tokens": 1}, "latency_ms": 175.46679399674758, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 303, "compiled_prompt_sha256": "f5114197072d0456f7f044d5682ca0a5e97b2cc3e89232c198233ea48eea91a3", "compiled_ids_sha256": "07c031aafebb4d50147cbd8c37e2b93c77bfe47f2a5860848d67bcd2dd338a9c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.625, 14.0, 14.6875, 17.125, 13.9375, 13.4375], "runtime_probabilities": [0.9650971293449402, 0.0012804720317944884, 0.0025465227663517, 0.029143409803509712, 0.001202892162837088, 0.0007295909454114735], "decide_seconds": 0.17189837200203328, "probabilities_f64_minus_runtime_max_abs": 1.4529434655408124e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.06659981351635547, "einstein": 0.002582355010837096, "darwin": 0.0027489025999948573, "sagan": 0.007019568700588639, "hypatia": 0.9193820664647633, "socrates": 0.0016672937074606541}, "confidence": 0.8131598918721201, "choice": "hypatia"}}, "usage": {"input_tokens": 303, "output_tokens": 1}, "latency_ms": 175.626801999897, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 303, "compiled_prompt_sha256": "316e4e233f9cec8ca26a21467e737810ce71f6e02d434c3980ecd0a7a99e5cdb", "compiled_ids_sha256": "5a8cdc19d6b5add526023fa46e680b4cf0175abacfd443699961c58b5f52d7dc", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 14.25, 14.3125, 15.25, 20.125, 13.8125], "runtime_probabilities": [0.06659982353448868, 0.0025823554024100304, 0.0027489028871059418, 0.0070195687003433704, 0.9193821549415588, 0.001667293719947338], "decide_seconds": 0.17199268100011977, "probabilities_f64_minus_runtime_max_abs": 8.847679555668009e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.0026342808336790725, "sagan": 0.0015977720919197137, "socrates": 0.0018105129743121707, "ibn_sina": 0.0529109449509066, "darwin": 0.0031775492270536164, "hypatia": 0.9378689399221288}, "confidence": 0.8485766374721595, "choice": "hypatia"}}, "usage": {"input_tokens": 303, "output_tokens": 1}, "latency_ms": 175.2348459995119, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 303, "compiled_prompt_sha256": "2b8889e768b472422cf59693e3e301d0243bcba7fc27cbe5e6a0f9fbb85d85ce", "compiled_ids_sha256": "f80051bc116833e78d71fc8244e831b2d2def084d69f32d6a13fd283ba03f22e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.625, 14.125, 14.25, 17.625, 14.8125, 20.5], "runtime_probabilities": [0.002634280826896429, 0.0015977721195667982, 0.0018105129711329937, 0.0529109463095665, 0.003177549224346876, 0.9378688931465149], "decide_seconds": 0.17167459299889742, "probabilities_f64_minus_runtime_max_abs": 4.6775613937732885e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.006939814557580157, "ibn_sina": 0.01217975385772193, "hypatia": 0.967557692243466, "einstein": 0.004480681042784462, "sagan": 0.007387394142461573, "darwin": 0.0014546641559857589}, "confidence": 0.8939118165893882, "choice": "hypatia"}}, "usage": {"input_tokens": 303, "output_tokens": 1}, "latency_ms": 175.94821999955457, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 303, "compiled_prompt_sha256": "b0c5bf7d8ff8c34974293c92ef2fbd6781c4f2e077caba91ae21a56bb5a9ee7b", "compiled_ids_sha256": "7efcb9b8a23db282a2a4c1f4e50da3b790d13754decda6c8aa3e8f85867e9ee4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.6875, 16.25, 20.625, 15.25, 15.75, 14.125], "runtime_probabilities": [0.006939813960343599, 0.012179753743112087, 0.9675576686859131, 0.004480681382119656, 0.007387393619865179, 0.0014546640450134873], "decide_seconds": 0.17226311899867142, "probabilities_f64_minus_runtime_max_abs": 2.355755290839312e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:30Z"} {"id": "c-hypatia-01-starter-02", "uid": "c052", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "prompt_sha256_by_order": ["44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d"], "request_sha256": "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "compiled_prompt_sha256_by_order": ["25078b32820a0c397cc2958b1cb9a27364e725d270a1e669793c995eab4ad072", "04867182226b2f863c7c6e001eb7a58b3bdf7b4d7644608ae220fb5bf0b7ff68", "ac59f35e73dad07af18bbc91f9c6c3bb5ff9b3bb081a4434c68a9758c0c4dc98", "3939c699cb8a8b20c876788fd495fa2bb581079d01287b27cc86cf7f299af77b", "5b257bb68c107e50dc9e2333fab374b26a415d3d36cb8718177d6c0312b8ca47", "eba7e53d2e352a92a741c6b4cdd60bc167e7ccc62b81d6aa2c2351260e3c586c"], "compiled_tokens_by_order": [273, 273, 273, 273, 273, 273], "choice": "ibn_sina", "probs": {"darwin": 0.024244332464149504, "einstein": 0.007394103640467734, "hypatia": 0.09588819499170538, "ibn_sina": 0.8028620269871073, "sagan": 0.05815913016697269, "socrates": 0.011452211749597372}, "confidence": 0.5846599557955062, "seconds": 0.17256244649979635, "seconds_all_orders": 1.035374678998778, "seconds_by_order": [0.17293441299989354, 0.1723391680025088, 0.1726886159995047, 0.17248571699747117, 0.17232429899740964, 0.17260246600199025], "argmax_stable": false, "spread": {"darwin": 0.04239765690904907, "einstein": 0.09902370834086464, "hypatia": 0.8319011599587908, "ibn_sina": 0.771114821071107, "sagan": 0.08317408660343603, "socrates": 0.06319871549932973}, "spread_2dp": {"darwin": 0.04, "einstein": 0.1, "hypatia": 0.8300000000000001, "ibn_sina": 0.77, "sagan": 0.08, "socrates": 0.060000000000000005}, "choices_by_order": ["ibn_sina", "ibn_sina", "hypatia", "hypatia", "ibn_sina", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.024244332464149504, "einstein": 0.007394103640467734, "hypatia": 0.09588819499170538, "ibn_sina": 0.8028620269871073, "sagan": 0.05815913016697269, "socrates": 0.011452211749597372}, {"ibn_sina": 0.45618758959066097, "hypatia": 0.3135328394431574, "sagan": 0.08982866247366042, "socrates": 0.06995863267674365, "darwin": 0.037446157663789115, "einstein": 0.0330461181519883}, {"hypatia": 0.5404342422610489, "darwin": 0.006803063114264118, "sagan": 0.050268214995863425, "ibn_sina": 0.2892736045332271, "einstein": 0.10641781198133238, "socrates": 0.006803063114264118}, {"ibn_sina": 0.4182212664805294, "einstein": 0.026735991077260744, "darwin": 0.044080197182328626, "sagan": 0.030295846930406482, "hypatia": 0.4739067811520608, "socrates": 0.006759917177413915}, {"einstein": 0.04379493278903145, "sagan": 0.00917991661209798, "socrates": 0.03410752795065723, "ibn_sina": 0.4708398998982048, "darwin": 0.026562969476601687, "hypatia": 0.41551475327340676}, {"socrates": 0.021819514257636393, "ibn_sina": 0.03174720591600039, "hypatia": 0.9277893549504962, "einstein": 0.010306808732363124, "sagan": 0.0066545758702243815, "darwin": 0.0016825402732795517}], "prompt_tokens": 273, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.024244332464149504, "einstein": 0.007394103640467734, "hypatia": 0.09588819499170538, "ibn_sina": 0.8028620269871073, "sagan": 0.05815913016697269, "socrates": 0.011452211749597372}, "confidence": 0.5846599557955062, "choice": "ibn_sina"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 172.9100929987908, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 273, "compiled_prompt_sha256": "25078b32820a0c397cc2958b1cb9a27364e725d270a1e669793c995eab4ad072", "compiled_ids_sha256": "9403e074d597096fdde8ed17e223082b95d8f4cb2f6a7e59d08436d83892a5ca", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 15.5625, 18.125, 20.25, 17.625, 16.0], "runtime_probabilities": [0.024244334548711777, 0.007394103799015284, 0.09588820487260818, 0.8028620481491089, 0.05815913528203964, 0.011452211998403072], "decide_seconds": 0.16934062200016342, "probabilities_f64_minus_runtime_max_abs": 2.1162001595520508e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.45618758959066097, "hypatia": 0.3135328394431574, "sagan": 0.08982866247366042, "socrates": 0.06995863267674365, "darwin": 0.037446157663789115, "einstein": 0.0330461181519883}, "confidence": 0.24100780343333206, "choice": "ibn_sina"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 172.3169479992066, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 273, "compiled_prompt_sha256": "04867182226b2f863c7c6e001eb7a58b3bdf7b4d7644608ae220fb5bf0b7ff68", "compiled_ids_sha256": "b2253b514a00fbf14563cf99c55e250e87a609c9d73ef595c32edd686b839250", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.25, 18.875, 17.625, 17.375, 16.75, 16.625], "runtime_probabilities": [0.45618757605552673, 0.31353282928466797, 0.08982865512371063, 0.0699586346745491, 0.03744615614414215, 0.033046118915081024], "decide_seconds": 0.16898015499828034, "probabilities_f64_minus_runtime_max_abs": 1.3535134235631574e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.5404342422610489, "darwin": 0.006803063114264118, "sagan": 0.050268214995863425, "ibn_sina": 0.2892736045332271, "einstein": 0.10641781198133238, "socrates": 0.006803063114264118}, "confidence": 0.3592772410973676, "choice": "hypatia"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 172.6666060021671, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 273, "compiled_prompt_sha256": "ac59f35e73dad07af18bbc91f9c6c3bb5ff9b3bb081a4434c68a9758c0c4dc98", "compiled_ids_sha256": "4993b0df859d5873143dcbb41ea55058d7b3c8fee204d3886f6b78a775d9f913", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.75, 15.375, 17.375, 19.125, 18.125, 15.375], "runtime_probabilities": [0.5404342412948608, 0.006803062744438648, 0.05026821047067642, 0.28927358984947205, 0.10641779750585556, 0.006803062744438648], "decide_seconds": 0.16930977199808694, "probabilities_f64_minus_runtime_max_abs": 1.4683755045208358e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.4182212664805294, "einstein": 0.026735991077260744, "darwin": 0.044080197182328626, "sagan": 0.030295846930406482, "hypatia": 0.4739067811520608, "socrates": 0.006759917177413915}, "confidence": 0.39019606835802056, "choice": "hypatia"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 172.46395800248138, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 273, "compiled_prompt_sha256": "3939c699cb8a8b20c876788fd495fa2bb581079d01287b27cc86cf7f299af77b", "compiled_ids_sha256": "310ec05141b06cb9f4723f00de0b2705dcadfcc935d5699672e3d80f9a6efa76", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.25, 16.5, 17.0, 16.625, 19.375, 15.125], "runtime_probabilities": [0.41822126507759094, 0.02673599310219288, 0.04408019781112671, 0.030295846983790398, 0.4739067852497101, 0.00675991689786315], "decide_seconds": 0.1690829040016979, "probabilities_f64_minus_runtime_max_abs": 4.097649275447424e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.04379493278903145, "sagan": 0.00917991661209798, "socrates": 0.03410752795065723, "ibn_sina": 0.4708398998982048, "darwin": 0.026562969476601687, "hypatia": 0.41551475327340676}, "confidence": 0.37980727523199687, "choice": "ibn_sina"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 172.30277899943758, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 273, "compiled_prompt_sha256": "5b257bb68c107e50dc9e2333fab374b26a415d3d36cb8718177d6c0312b8ca47", "compiled_ids_sha256": "0d8258aa2299d8d5fe7cb95bafb958f54ed427600a110a229e2ae0f5e3e571c1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 15.5625, 16.875, 19.5, 16.625, 19.375], "runtime_probabilities": [0.04379492998123169, 0.0091799171641469, 0.03410752862691879, 0.47083988785743713, 0.026562971994280815, 0.41551473736763], "decide_seconds": 0.16897453400088125, "probabilities_f64_minus_runtime_max_abs": 1.590577675125715e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.021819514257636393, "ibn_sina": 0.03174720591600039, "hypatia": 0.9277893549504962, "einstein": 0.010306808732363124, "sagan": 0.0066545758702243815, "darwin": 0.0016825402732795517}, "confidence": 0.8025520680649427, "choice": "hypatia"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 172.58034699989366, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 273, "compiled_prompt_sha256": "eba7e53d2e352a92a741c6b4cdd60bc167e7ccc62b81d6aa2c2351260e3c586c", "compiled_ids_sha256": "4ee5763d44578af9946676e2779063aa20d3ae5a4788dd98641c203c5f718886", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 17.125, 20.5, 16.0, 15.5625, 14.1875], "runtime_probabilities": [0.021819517016410828, 0.031747207045555115, 0.9277893900871277, 0.010306810028851032, 0.0066545759327709675, 0.0016825402854010463], "decide_seconds": 0.1691753230006725, "probabilities_f64_minus_runtime_max_abs": 3.513663149323776e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:31Z"} {"id": "c-hypatia-01-starter-02", "uid": "c053", "task": "kperm", "label": "hypatia", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "prompt_sha256_by_order": ["a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131"], "request_sha256": "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "compiled_prompt_sha256_by_order": ["ae3c350b97daa5c6854c06d3ed1c5bfc8800493c92943b74514bed4d49857d5a", "24ffa2e91cc6fc8072e90642b47278ef16da2be9bbf05d8b4872fa44902df1e0", "0c5cc6b3a9d65f22fbf40b8e5f4b5bf10ea3acf40acf416024d37c4df7f92aff", "945fb4ad4fb6bac54e3c50412746bb58a5e332203b5e076b17b8e5612271476d", "f989215095bc58534df29645ac95b12f450bfaefa5e7431dad9bb9f9b043f7ea", "8abc5ce7d1a3f882b77d7861f043a3123a08448019880594565110d7eddd71e1"], "compiled_tokens_by_order": [309, 309, 309, 309, 309, 309], "choice": "hypatia", "probs": {"darwin": 0.008184577177092757, "einstein": 0.006785252799804639, "hypatia": 0.6501814977513486, "ibn_sina": 0.04156471254741746, "sagan": 0.27103597231022475, "socrates": 0.02224798741411187}, "confidence": 0.48440882971543786, "seconds": 0.1768138618329734, "seconds_all_orders": 1.0608831709978404, "seconds_by_order": [0.17686784299803548, 0.1774707180011319, 0.1769502620009007, 0.17660339500071132, 0.17690462300015497, 0.176086329996906], "argmax_stable": false, "spread": {"darwin": 0.005322472945614616, "einstein": 0.016695128196512792, "hypatia": 0.8416819637310433, "ibn_sina": 0.03453738183759986, "sagan": 0.828497565589049, "socrates": 0.041593523730208616}, "spread_2dp": {"darwin": 0.0, "einstein": 0.01, "hypatia": 0.84, "ibn_sina": 0.03, "sagan": 0.8300000000000001, "socrates": 0.04}, "choices_by_order": ["hypatia", "sagan", "sagan", "hypatia", "sagan", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.008184577177092757, "einstein": 0.006785252799804639, "hypatia": 0.6501814977513486, "ibn_sina": 0.04156471254741746, "sagan": 0.27103597231022475, "socrates": 0.02224798741411187}, {"ibn_sina": 0.017251853233027786, "hypatia": 0.18547490159253574, "sagan": 0.7335674659746229, "socrates": 0.046895399150581864, "darwin": 0.010463777922693871, "einstein": 0.006346602126538003}, {"hypatia": 0.049956384235244886, "darwin": 0.005265364160309355, "sagan": 0.8854980981444158, "ibn_sina": 0.030300078687060878, "einstein": 0.02082491922813978, "socrates": 0.008155155544829401}, {"ibn_sina": 0.030510186481928894, "einstein": 0.008741314780171556, "darwin": 0.006807742795871253, "sagan": 0.0570005325553669, "hypatia": 0.8916383479662882, "socrates": 0.005301875420373247}, {"einstein": 0.017102078569828476, "sagan": 0.499795997373644, "socrates": 0.017102078569828476, "ibn_sina": 0.019379193875628458, "darwin": 0.005552232004651172, "hypatia": 0.4410684196064194}, {"socrates": 0.009023271242876107, "ibn_sina": 0.007027330709817601, "hypatia": 0.8646345514906647, "einstein": 0.023041730323050794, "sagan": 0.09113181125651142, "darwin": 0.005141304977079255}], "prompt_tokens": 309, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.008184577177092757, "einstein": 0.006785252799804639, "hypatia": 0.6501814977513486, "ibn_sina": 0.04156471254741746, "sagan": 0.27103597231022475, "socrates": 0.02224798741411187}, "confidence": 0.48440882971543786, "choice": "hypatia"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 176.84081299739773, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 309, "compiled_prompt_sha256": "ae3c350b97daa5c6854c06d3ed1c5bfc8800493c92943b74514bed4d49857d5a", "compiled_ids_sha256": "55a7c5fbd189b1b2d5df695ec9fc001b672fd7fe50d6a7ee300faac082ac3ee1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.375, 15.1875, 19.75, 17.0, 18.875, 16.375], "runtime_probabilities": [0.008184577338397503, 0.006785253062844276, 0.6501815319061279, 0.041564710438251495, 0.27103596925735474, 0.022247986868023872], "decide_seconds": 0.17288217399982386, "probabilities_f64_minus_runtime_max_abs": 3.4154779338102514e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.017251853233027786, "hypatia": 0.18547490159253574, "sagan": 0.7335674659746229, "socrates": 0.046895399150581864, "darwin": 0.010463777922693871, "einstein": 0.006346602126538003}, "confidence": 0.5350168786406414, "choice": "sagan"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 177.44608800057904, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 309, "compiled_prompt_sha256": "24ffa2e91cc6fc8072e90642b47278ef16da2be9bbf05d8b4872fa44902df1e0", "compiled_ids_sha256": "687ac4306df3483e9b2d71c057b95ff6abde6d6646c2e346aba92749337d1d43", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.875, 18.25, 19.625, 16.875, 15.375, 14.875], "runtime_probabilities": [0.01725185289978981, 0.18547490239143372, 0.733567476272583, 0.04689539596438408, 0.010463777929544449, 0.006346601992845535], "decide_seconds": 0.17323797200151603, "probabilities_f64_minus_runtime_max_abs": 1.0297960151284258e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.049956384235244886, "darwin": 0.005265364160309355, "sagan": 0.8854980981444158, "ibn_sina": 0.030300078687060878, "einstein": 0.02082491922813978, "socrates": 0.008155155544829401}, "confidence": 0.7149181692081652, "choice": "sagan"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 176.92658200030564, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 309, "compiled_prompt_sha256": "0c5cc6b3a9d65f22fbf40b8e5f4b5bf10ea3acf40acf416024d37c4df7f92aff", "compiled_ids_sha256": "badc51827791db05c8cc51b2e7e85e2764f97366133d3f071b176052120a138b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.25, 15.0, 20.125, 16.75, 16.375, 15.4375], "runtime_probabilities": [0.049956388771533966, 0.005265363957732916, 0.8854981064796448, 0.03030007891356945, 0.020824920386075974, 0.008155155926942825], "decide_seconds": 0.17298067400042783, "probabilities_f64_minus_runtime_max_abs": 8.335228951494855e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.030510186481928894, "einstein": 0.008741314780171556, "darwin": 0.006807742795871253, "sagan": 0.0570005325553669, "hypatia": 0.8916383479662882, "socrates": 0.005301875420373247}, "confidence": 0.7347822301501763, "choice": "hypatia"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 176.58135499732452, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 309, "compiled_prompt_sha256": "945fb4ad4fb6bac54e3c50412746bb58a5e332203b5e076b17b8e5612271476d", "compiled_ids_sha256": "4ee8851717638c4aadbd737007371dc1d5c01ae881489f44217879599e0ffb6a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 15.375, 15.125, 17.25, 20.0, 14.875], "runtime_probabilities": [0.03051018714904785, 0.008741315454244614, 0.006807742640376091, 0.057000532746315, 0.8916383981704712, 0.005301875527948141], "decide_seconds": 0.1728804949998448, "probabilities_f64_minus_runtime_max_abs": 5.020418303303131e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.017102078569828476, "sagan": 0.499795997373644, "socrates": 0.017102078569828476, "ibn_sina": 0.019379193875628458, "darwin": 0.005552232004651172, "hypatia": 0.4410684196064194}, "confidence": 0.46862550898446953, "choice": "sagan"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 176.88262299998314, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 309, "compiled_prompt_sha256": "f989215095bc58534df29645ac95b12f450bfaefa5e7431dad9bb9f9b043f7ea", "compiled_ids_sha256": "e5105ce9454b3cc1b9735f46cb49895d47e8684307c9d484e22722a1e07348ce", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 19.375, 16.0, 16.125, 14.875, 19.25], "runtime_probabilities": [0.01710207760334015, 0.4997960031032562, 0.01710207760334015, 0.01937919296324253, 0.005552232265472412, 0.4410684108734131], "decide_seconds": 0.1731841920009174, "probabilities_f64_minus_runtime_max_abs": 8.733006318095704e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.009023271242876107, "ibn_sina": 0.007027330709817601, "hypatia": 0.8646345514906647, "einstein": 0.023041730323050794, "sagan": 0.09113181125651142, "darwin": 0.005141304977079255}, "confidence": 0.7012109637570246, "choice": "hypatia"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 176.0640100001183, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 309, "compiled_prompt_sha256": "8abc5ce7d1a3f882b77d7861f043a3123a08448019880594565110d7eddd71e1", "compiled_ids_sha256": "6285989cc94b20ae0b5f839c5aebba799057abbd03ed2deb13ff20954ec8fe1a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5625, 15.3125, 20.125, 16.5, 17.875, 15.0], "runtime_probabilities": [0.0090232715010643, 0.007027330808341503, 0.8646345138549805, 0.023041728883981705, 0.09113180637359619, 0.005141304340213537], "decide_seconds": 0.17236737799976254, "probabilities_f64_minus_runtime_max_abs": 3.763568423575947e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:33Z"} {"id": "c-ibn_sina-01-dessert-01", "uid": "c054", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "prompt_sha256_by_order": ["2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d"], "request_sha256": "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "compiled_prompt_sha256_by_order": ["066735836834e07c7890804951ba5a2894b840544a1b10249ec2970aea3a1c62", "f783b341158d6a1bdbae00793d7a6096835f1bd52d229fe58519cbabe172e45f", "ede7ee9b39d29cf1e7ea56a722fe5f616f8cf1b62d53cb01af5bacb7cde9fb6c", "ff151f825d9b8ab7de0b22957e316fe6591121215d83ef608f868eb052250c79", "aca7682169f325099be485818280caadfbc9d8403c61689472e66a9ae61b1114", "3a0cf905adedfda68784b2d3929d9f06bdb65d705c7652ee9ac3948298c869eb"], "compiled_tokens_by_order": [311, 311, 311, 311, 311, 311], "choice": "sagan", "probs": {"darwin": 0.14399520332504923, "einstein": 0.060026131282097495, "hypatia": 0.3048378478312747, "ibn_sina": 0.052972874930587326, "sagan": 0.39141954458374684, "socrates": 0.046748398047244476}, "confidence": 0.1762203171232961, "seconds": 0.17668479116764502, "seconds_all_orders": 1.06010874700587, "seconds_by_order": [0.17677325399927213, 0.17628864800281008, 0.1768875929992646, 0.17673974400167936, 0.17633544700220227, 0.17708406100064167], "argmax_stable": false, "spread": {"darwin": 0.35262978504245657, "einstein": 0.09936702913025167, "hypatia": 0.29223731697414607, "ibn_sina": 0.07118572898226683, "sagan": 0.5188816170221898, "socrates": 0.08733182282331794}, "spread_2dp": {"darwin": 0.35, "einstein": 0.09999999999999999, "hypatia": 0.29, "ibn_sina": 0.06999999999999999, "sagan": 0.52, "socrates": 0.09}, "choices_by_order": ["sagan", "sagan", "darwin", "sagan", "hypatia", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.14399520332504923, "einstein": 0.060026131282097495, "hypatia": 0.3048378478312747, "ibn_sina": 0.052972874930587326, "sagan": 0.39141954458374684, "socrates": 0.046748398047244476}, {"ibn_sina": 0.07030343591285264, "hypatia": 0.06204256443414706, "sagan": 0.58864346196144, "socrates": 0.06204256443414706, "darwin": 0.1686491754923414, "einstein": 0.04831879776507183}, {"hypatia": 0.07172579335582754, "darwin": 0.4677109247355268, "sagan": 0.2836810157346449, "ibn_sina": 0.05586010403193626, "einstein": 0.07172579335582754, "socrates": 0.04929636878623701}, {"ibn_sina": 0.09118383244350318, "einstein": 0.07101404011045206, "darwin": 0.1930361747977036, "sagan": 0.4086575852535857, "hypatia": 0.1930361747977036, "socrates": 0.043072192597051896}, {"einstein": 0.1476858268953235, "sagan": 0.06976184493925021, "socrates": 0.1150178376346259, "ibn_sina": 0.06976184493925021, "darwin": 0.24349276418325697, "hypatia": 0.35427988140829314}, {"socrates": 0.13040401542036983, "ibn_sina": 0.019998103461236358, "hypatia": 0.18973672286837764, "einstein": 0.0896252817097128, "sagan": 0.4551547368472332, "darwin": 0.11508113969307018}], "prompt_tokens": 311, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.14399520332504923, "einstein": 0.060026131282097495, "hypatia": 0.3048378478312747, "ibn_sina": 0.052972874930587326, "sagan": 0.39141954458374684, "socrates": 0.046748398047244476}, "confidence": 0.1762203171232961, "choice": "sagan"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 176.74745400290703, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 311, "compiled_prompt_sha256": "066735836834e07c7890804951ba5a2894b840544a1b10249ec2970aea3a1c62", "compiled_ids_sha256": "8b485c39edfb00d587270bcc96d96b07d12859a38e102a757f143839cbda8eb7", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 17.25, 18.875, 17.125, 19.125, 17.0], "runtime_probabilities": [0.14399521052837372, 0.060026127845048904, 0.30483782291412354, 0.05297287181019783, 0.39141952991485596, 0.04674839600920677], "decide_seconds": 0.1727866549990722, "probabilities_f64_minus_runtime_max_abs": 2.491715117569626e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.07030343591285264, "hypatia": 0.06204256443414706, "sagan": 0.58864346196144, "socrates": 0.06204256443414706, "darwin": 0.1686491754923414, "einstein": 0.04831879776507183}, "confidence": 0.27996506587055847, "choice": "sagan"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 176.26592800297658, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 311, "compiled_prompt_sha256": "f783b341158d6a1bdbae00793d7a6096835f1bd52d229fe58519cbabe172e45f", "compiled_ids_sha256": "ed960119ce5f8e6565ff801303fdbcec4f6bdb265f371600327c44bafaa8469c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 17.25, 19.5, 17.25, 18.25, 17.0], "runtime_probabilities": [0.07030343264341354, 0.062042560428380966, 0.5886434316635132, 0.062042560428380966, 0.16864916682243347, 0.04831879585981369], "decide_seconds": 0.1725666660022398, "probabilities_f64_minus_runtime_max_abs": 3.02979268340664e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.07172579335582754, "darwin": 0.4677109247355268, "sagan": 0.2836810157346449, "ibn_sina": 0.05586010403193626, "einstein": 0.07172579335582754, "socrates": 0.04929636878623701}, "confidence": 0.2184572405303642, "choice": "darwin"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 176.8658030014194, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 311, "compiled_prompt_sha256": "ede7ee9b39d29cf1e7ea56a722fe5f616f8cf1b62d53cb01af5bacb7cde9fb6c", "compiled_ids_sha256": "3e0522a64fe8a366e627d965d820c44e3c949ca7db48f92f8d91ac37b00bbdf3", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 19.25, 18.75, 17.125, 17.375, 17.0], "runtime_probabilities": [0.07172580063343048, 0.467710942029953, 0.2836810350418091, 0.05586010590195656, 0.07172580063343048, 0.04929636791348457], "decide_seconds": 0.17314078199706273, "probabilities_f64_minus_runtime_max_abs": 1.930716420073253e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.09118383244350318, "einstein": 0.07101404011045206, "darwin": 0.1930361747977036, "sagan": 0.4086575852535857, "hypatia": 0.1930361747977036, "socrates": 0.043072192597051896}, "confidence": 0.1391731343723105, "choice": "sagan"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 176.71761400197283, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 311, "compiled_prompt_sha256": "ff151f825d9b8ab7de0b22957e316fe6591121215d83ef608f868eb052250c79", "compiled_ids_sha256": "c5abb226e29b161b96dee1f321148a8bfae64ee1d02ae5e8dc7ab6a26bf07807", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 17.125, 18.125, 18.875, 18.125, 16.625], "runtime_probabilities": [0.09118383377790451, 0.07101403176784515, 0.19303618371486664, 0.40865758061408997, 0.19303618371486664, 0.04307219013571739], "decide_seconds": 0.17302887299956637, "probabilities_f64_minus_runtime_max_abs": 8.917163035748743e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.1476858268953235, "sagan": 0.06976184493925021, "socrates": 0.1150178376346259, "ibn_sina": 0.06976184493925021, "darwin": 0.24349276418325697, "hypatia": 0.35427988140829314}, "confidence": 0.09902862019196501, "choice": "hypatia"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 176.31324700050754, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 311, "compiled_prompt_sha256": "aca7682169f325099be485818280caadfbc9d8403c61689472e66a9ae61b1114", "compiled_ids_sha256": "e83fedc7e71580c6254aa4d6e7eeadd714de9f75b82ed69cb74b9b88884c3021", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 17.375, 17.875, 17.375, 18.625, 19.0], "runtime_probabilities": [0.14768582582473755, 0.06976184248924255, 0.1150178387761116, 0.06976184248924255, 0.24349278211593628, 0.35427987575531006], "decide_seconds": 0.17259881700010737, "probabilities_f64_minus_runtime_max_abs": 1.7932679313714672e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.13040401542036983, "ibn_sina": 0.019998103461236358, "hypatia": 0.18973672286837764, "einstein": 0.0896252817097128, "sagan": 0.4551547368472332, "darwin": 0.11508113969307018}, "confidence": 0.17259306436967758, "choice": "sagan"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 177.06149099831237, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 311, "compiled_prompt_sha256": "3a0cf905adedfda68784b2d3929d9f06bdb65d705c7652ee9ac3948298c869eb", "compiled_ids_sha256": "84a654325c4190581d7415981d6d5c64d1878d9d2ff32034c6ae1e2cbe18310d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.0, 16.125, 18.375, 17.625, 19.25, 17.875], "runtime_probabilities": [0.1304040253162384, 0.019998105242848396, 0.1897367388010025, 0.089625284075737, 0.45515477657318115, 0.11508114635944366], "decide_seconds": 0.17331695100074285, "probabilities_f64_minus_runtime_max_abs": 3.972594797563289e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:34Z"} {"id": "c-ibn_sina-01-dessert-02", "uid": "c055", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "prompt_sha256_by_order": ["a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e"], "request_sha256": "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "compiled_prompt_sha256_by_order": ["aa541bf54e4da10cb7233c63ec2579c3ba492ff5efbefd219bec4731f7f5d42b", "719ff064b126def9a53c8d0e1237002376a4bec9812d53dafb458ae3d6e15ff6", "38403e234882e32610254e467c79c997b2faac2df87dc46093553a8908056822", "61119adef9837a7be193e2b75f6706c60b7e71f5a804a76e053a6ca7453326c2", "2b1dbb917cf71dfd3a50147a103cc32e984ad22b21f79511951e9458a5e6fa2b", "cb95ff037757ccada52a6794bc80b016c20ce88676cada73060577fd944f2636"], "compiled_tokens_by_order": [323, 323, 323, 323, 323, 323], "choice": "ibn_sina", "probs": {"darwin": 0.003341432906484987, "einstein": 0.0022965310124264888, "hypatia": 0.004290484780083517, "ibn_sina": 0.9862399962665045, "sagan": 0.0026023105641510683, "socrates": 0.0012292444703495653}, "confidence": 0.9476581689858663, "seconds": 0.18398548116723154, "seconds_all_orders": 1.1039128870033892, "seconds_by_order": [0.18430060500031686, 0.18405659699783428, 0.1835262420026993, 0.18407817700062878, 0.18410600700008217, 0.1838452590018278], "argmax_stable": true, "spread": {"darwin": 0.005180046077777531, "einstein": 0.0038524079727360917, "hypatia": 0.005200192065622473, "ibn_sina": 0.013386195743482543, "sagan": 0.0021732022075228734, "socrates": 0.0021975321812879043}, "spread_2dp": {"darwin": 0.01, "einstein": 0.0, "hypatia": 0.01, "ibn_sina": 0.020000000000000018, "sagan": 0.0, "socrates": 0.0}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.003341432906484987, "einstein": 0.0022965310124264888, "hypatia": 0.004290484780083517, "ibn_sina": 0.9862399962665045, "sagan": 0.0026023105641510683, "socrates": 0.0012292444703495653}, {"ibn_sina": 0.996206123627101, "hypatia": 0.0010293774351501995, "sagan": 0.00042910835662819495, "socrates": 0.0006646160241428493, "darwin": 0.0012416662003495544, "einstein": 0.00042910835662819495}, {"hypatia": 0.006229569500772672, "darwin": 0.003549501707007859, "sagan": 0.0007440159765011069, "ibn_sina": 0.9841784472208328, "einstein": 0.004281516329364287, "socrates": 0.0010169492655211847}, {"ibn_sina": 0.99530499034753, "einstein": 0.0011653823296696793, "darwin": 0.0006640148356906936, "sagan": 0.0010284462962603118, "hypatia": 0.001592888784166027, "socrates": 0.00024427740668343947}, {"einstein": 0.004016560410428086, "sagan": 0.0010810425689169116, "socrates": 0.0010810425689169116, "ibn_sina": 0.9828199278836185, "darwin": 0.0058440609134682246, "hypatia": 0.00515736565465141}, {"socrates": 0.0024418095879713436, "ibn_sina": 0.9850962960142717, "hypatia": 0.004856118221183908, "einstein": 0.004025863406687869, "sagan": 0.0016782295506617055, "darwin": 0.0019016832192233469}], "prompt_tokens": 323, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.003341432906484987, "einstein": 0.0022965310124264888, "hypatia": 0.004290484780083517, "ibn_sina": 0.9862399962665045, "sagan": 0.0026023105641510683, "socrates": 0.0012292444703495653}, "confidence": 0.9476581689858663, "choice": "ibn_sina"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 184.27402599991183, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "aa541bf54e4da10cb7233c63ec2579c3ba492ff5efbefd219bec4731f7f5d42b", "compiled_ids_sha256": "6d3e88b7607381a1857be3378d363a3a49ad70080bd24153eee4a6f79fe0c29c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.3125, 14.9375, 15.5625, 21.0, 15.0625, 14.3125], "runtime_probabilities": [0.0033414328936487436, 0.0022965308744460344, 0.004290484823286533, 0.9862399697303772, 0.0026023106183856726, 0.0012292445171624422], "decide_seconds": 0.18017889699694933, "probabilities_f64_minus_runtime_max_abs": 2.65361272955289e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.996206123627101, "hypatia": 0.0010293774351501995, "sagan": 0.00042910835662819495, "socrates": 0.0006646160241428493, "darwin": 0.0012416662003495544, "einstein": 0.00042910835662819495}, "confidence": 0.9828699729572516, "choice": "ibn_sina"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 184.03370799933327, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "719ff064b126def9a53c8d0e1237002376a4bec9812d53dafb458ae3d6e15ff6", "compiled_ids_sha256": "2e5335910a403d03f98920ce9cac22c9f0273eeda65f8ae3961984e75c23f585", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.625, 13.75, 12.875, 13.3125, 13.9375, 12.875], "runtime_probabilities": [0.9962060451507568, 0.0010293773375451565, 0.00042910833144560456, 0.0006646160036325455, 0.001241666148416698, 0.00042910833144560456], "decide_seconds": 0.18023705700034043, "probabilities_f64_minus_runtime_max_abs": 7.847634420077299e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.006229569500772672, "darwin": 0.003549501707007859, "sagan": 0.0007440159765011069, "ibn_sina": 0.9841784472208328, "einstein": 0.004281516329364287, "socrates": 0.0010169492655211847}, "confidence": 0.9424749358872395, "choice": "ibn_sina"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 183.50396199821262, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "38403e234882e32610254e467c79c997b2faac2df87dc46093553a8908056822", "compiled_ids_sha256": "b3ae8862f8055606217f15284cb54126369ae13606edd1ab769fae010e5b8581", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 15.375, 13.8125, 21.0, 15.5625, 14.125], "runtime_probabilities": [0.006229569669812918, 0.0035495019983500242, 0.0007440160261467099, 0.9841784834861755, 0.004281516652554274, 0.0010169493034482002], "decide_seconds": 0.1797184909992211, "probabilities_f64_minus_runtime_max_abs": 3.626534272882509e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.99530499034753, "einstein": 0.0011653823296696793, "darwin": 0.0006640148356906936, "sagan": 0.0010284462962603118, "hypatia": 0.001592888784166027, "socrates": 0.00024427740668343947}, "confidence": 0.9794708210277864, "choice": "ibn_sina"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 184.05526700007613, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "61119adef9837a7be193e2b75f6706c60b7e71f5a804a76e053a6ca7453326c2", "compiled_ids_sha256": "f81a84892bebe7604274c8a613f81476a0366788d4ddfa128180c78947b642c4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.625, 13.875, 13.3125, 13.75, 14.1875, 12.3125], "runtime_probabilities": [0.995305061340332, 0.001165382331237197, 0.0006640148931182921, 0.0010284463642165065, 0.0015928889624774456, 0.0002442774421069771], "decide_seconds": 0.18027300700123305, "probabilities_f64_minus_runtime_max_abs": 7.099280208144876e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.004016560410428086, "sagan": 0.0010810425689169116, "socrates": 0.0010810425689169116, "ibn_sina": 0.9828199278836185, "darwin": 0.0058440609134682246, "hypatia": 0.00515736565465141}, "confidence": 0.9379511408001582, "choice": "ibn_sina"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 184.08367699885275, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "2b1dbb917cf71dfd3a50147a103cc32e984ad22b21f79511951e9458a5e6fa2b", "compiled_ids_sha256": "486109abcac25cd8a2cd9f0d4081419470451170e163118482ebcf0a6ddca8fd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.25, 13.9375, 13.9375, 20.75, 15.625, 15.5], "runtime_probabilities": [0.00401656050235033, 0.001081042573787272, 0.001081042573787272, 0.9828199148178101, 0.00584406079724431, 0.005157365929335356], "decide_seconds": 0.1802728559996467, "probabilities_f64_minus_runtime_max_abs": 1.3065808435186455e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0024418095879713436, "ibn_sina": 0.9850962960142717, "hypatia": 0.004856118221183908, "einstein": 0.004025863406687869, "sagan": 0.0016782295506617055, "darwin": 0.0019016832192233469}, "confidence": 0.9440821440163067, "choice": "ibn_sina"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 183.8230490029673, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "cb95ff037757ccada52a6794bc80b016c20ce88676cada73060577fd944f2636", "compiled_ids_sha256": "ee9ff1d8debac08d3d0dc0979ef0b33d378160d30a736188f3070566f29f0134", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.125, 21.125, 15.8125, 15.625, 14.75, 14.875], "runtime_probabilities": [0.002441809745505452, 0.985096275806427, 0.0048561180010437965, 0.004025863483548164, 0.0016782296588644385, 0.0019016831647604704], "decide_seconds": 0.17997762899904046, "probabilities_f64_minus_runtime_max_abs": 2.0207844730890656e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:35Z"} {"id": "c-ibn_sina-01-dessert-02", "uid": "c056", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "prompt_sha256_by_order": ["ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595"], "request_sha256": "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "compiled_prompt_sha256_by_order": ["725690d1f22c93d0340b285452fdb898cbeffbf60adbcee2a6cf9213e82abbc0", "b8aac5aa144a3a3b9b2042865a2d2015541abec27b5967e49c0db4498c4881a8", "ef9e5a25d1357e0028a89e6361af2e9194a3bac6232a11e3d3d52552c9527f95", "e919be9065d309bf82ab619c6d58503f847ac29334d5c3dfc4aee4836de71670", "2a5d84bd4af93d07480a033a6331117b9c81bae0d818e013fc68eebed536bfd5", "ebf830a50a0084bf727dee0ecca59607345f1d5b5ff40a355f83f10100363532"], "compiled_tokens_by_order": [342, 342, 342, 342, 342, 342], "choice": "hypatia", "probs": {"darwin": 0.03557693773770083, "einstein": 0.004814814945423077, "hypatia": 0.4911244763131812, "ibn_sina": 0.3824881267382392, "sagan": 0.04568169230312515, "socrates": 0.0403139519623305}, "confidence": 0.3684327621349519, "seconds": 0.18611782116628697, "seconds_all_orders": 1.116706926997722, "seconds_by_order": [0.18662629800019204, 0.18576810399827082, 0.18609376099993824, 0.18623509999815724, 0.18583399300041492, 0.18614967100074864], "argmax_stable": false, "spread": {"darwin": 0.129123204090873, "einstein": 0.061113500026623704, "hypatia": 0.7510955882070174, "ibn_sina": 0.7378230690327857, "sagan": 0.03618381800985897, "socrates": 0.15953684460194115}, "spread_2dp": {"darwin": 0.13, "einstein": 0.07, "hypatia": 0.75, "ibn_sina": 0.73, "sagan": 0.04, "socrates": 0.16}, "choices_by_order": ["hypatia", "ibn_sina", "ibn_sina", "ibn_sina", "hypatia", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.03557693773770083, "einstein": 0.004814814945423077, "hypatia": 0.4911244763131812, "ibn_sina": 0.3824881267382392, "sagan": 0.04568169230312515, "socrates": 0.0403139519623305}, {"ibn_sina": 0.553835612409774, "hypatia": 0.03540552616146392, "sagan": 0.04546159548252244, "socrates": 0.17980409807975575, "darwin": 0.14003157238396147, "einstein": 0.04546159548252244}, {"hypatia": 0.1328911308516797, "darwin": 0.010908368293088482, "sagan": 0.010247463668507657, "ibn_sina": 0.7647356572165036, "einstein": 0.04314343352103912, "socrates": 0.03807394644918143}, {"ibn_sina": 0.592295967101875, "einstein": 0.00957357265884363, "darwin": 0.022965806926298887, "sagan": 0.0378642143781813, "hypatia": 0.3170331854569865, "socrates": 0.020267253477814613}, {"einstein": 0.06592831497204678, "sagan": 0.009497874293266186, "socrates": 0.10869741523583125, "ibn_sina": 0.17921174056943917, "darwin": 0.08465363210350303, "hypatia": 0.5520110228259136}, {"socrates": 0.05697394963202081, "ibn_sina": 0.026912588183717872, "hypatia": 0.7865011143684814, "einstein": 0.030495957668404457, "sagan": 0.034556447256743936, "darwin": 0.06455994289063166}], "prompt_tokens": 342, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.03557693773770083, "einstein": 0.004814814945423077, "hypatia": 0.4911244763131812, "ibn_sina": 0.3824881267382392, "sagan": 0.04568169230312515, "socrates": 0.0403139519623305}, "confidence": 0.3684327621349519, "choice": "hypatia"}}, "usage": {"input_tokens": 342, "output_tokens": 1}, "latency_ms": 186.59934799870825, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 342, "compiled_prompt_sha256": "725690d1f22c93d0340b285452fdb898cbeffbf60adbcee2a6cf9213e82abbc0", "compiled_ids_sha256": "94b95ec2aa38d8983aa6f7c9626041fbc139ec74d98175507b3c7f676e48c4ab", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 14.5, 19.125, 18.875, 16.75, 16.625], "runtime_probabilities": [0.03557693585753441, 0.004814814776182175, 0.4911244511604309, 0.38248810172080994, 0.04568168893456459, 0.040313951671123505], "decide_seconds": 0.18231269100215286, "probabilities_f64_minus_runtime_max_abs": 2.5152750271484337e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.553835612409774, "hypatia": 0.03540552616146392, "sagan": 0.04546159548252244, "socrates": 0.17980409807975575, "darwin": 0.14003157238396147, "einstein": 0.04546159548252244}, "confidence": 0.2686610094721349, "choice": "ibn_sina"}}, "usage": {"input_tokens": 342, "output_tokens": 1}, "latency_ms": 185.74438399809878, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 342, "compiled_prompt_sha256": "b8aac5aa144a3a3b9b2042865a2d2015541abec27b5967e49c0db4498c4881a8", "compiled_ids_sha256": "3f3a2569cd4a2ce0f1e66875c36c96917d14aaaa302d580d85aae029b59bcd14", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.125, 16.375, 16.625, 18.0, 17.75, 16.625], "runtime_probabilities": [0.5538356304168701, 0.03540552407503128, 0.04546159505844116, 0.17980410158634186, 0.1400315761566162, 0.04546159505844116], "decide_seconds": 0.18173400500018033, "probabilities_f64_minus_runtime_max_abs": 1.800709614752094e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.1328911308516797, "darwin": 0.010908368293088482, "sagan": 0.010247463668507657, "ibn_sina": 0.7647356572165036, "einstein": 0.04314343352103912, "socrates": 0.03807394644918143}, "confidence": 0.5369933814960728, "choice": "ibn_sina"}}, "usage": {"input_tokens": 342, "output_tokens": 1}, "latency_ms": 186.07025100209285, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 342, "compiled_prompt_sha256": "ef9e5a25d1357e0028a89e6361af2e9194a3bac6232a11e3d3d52552c9527f95", "compiled_ids_sha256": "864494f70d217d88f3658a0a30829020dc43e406f70060ad983ecbf868de4275", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.875, 15.375, 15.3125, 19.625, 16.75, 16.625], "runtime_probabilities": [0.13289111852645874, 0.010908368043601513, 0.010247464291751385, 0.7647356390953064, 0.043143436312675476, 0.03807394579052925], "decide_seconds": 0.18207854199863505, "probabilities_f64_minus_runtime_max_abs": 1.8121197209453044e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.592295967101875, "einstein": 0.00957357265884363, "darwin": 0.022965806926298887, "sagan": 0.0378642143781813, "hypatia": 0.3170331854569865, "socrates": 0.020267253477814613}, "confidence": 0.4371158559017838, "choice": "ibn_sina"}}, "usage": {"input_tokens": 342, "output_tokens": 1}, "latency_ms": 186.21085000268067, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 342, "compiled_prompt_sha256": "e919be9065d309bf82ab619c6d58503f847ac29334d5c3dfc4aee4836de71670", "compiled_ids_sha256": "a58249ff3fdeebc3c38396447d68f6b88df0bc950b26791659c8521322bd44f1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.375, 15.25, 16.125, 16.625, 18.75, 16.0], "runtime_probabilities": [0.5922960042953491, 0.009573572315275669, 0.02296580746769905, 0.037864215672016144, 0.31703320145606995, 0.02026725374162197], "decide_seconds": 0.18210108199855313, "probabilities_f64_minus_runtime_max_abs": 3.7193474078378586e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.06592831497204678, "sagan": 0.009497874293266186, "socrates": 0.10869741523583125, "ibn_sina": 0.17921174056943917, "darwin": 0.08465363210350303, "hypatia": 0.5520110228259136}, "confidence": 0.2689631961001332, "choice": "hypatia"}}, "usage": {"input_tokens": 342, "output_tokens": 1}, "latency_ms": 185.81113399704918, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 342, "compiled_prompt_sha256": "2a5d84bd4af93d07480a033a6331117b9c81bae0d818e013fc68eebed536bfd5", "compiled_ids_sha256": "7855a162de97f7e775b36190545344eebc7e1d0de7bb34c8275ef5493cc8b43c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 15.1875, 17.625, 18.125, 17.375, 19.25], "runtime_probabilities": [0.06592831760644913, 0.009497874416410923, 0.10869740694761276, 0.17921173572540283, 0.08465363830327988, 0.5520110130310059], "decide_seconds": 0.18178058499688632, "probabilities_f64_minus_runtime_max_abs": 9.79490777375247e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.05697394963202081, "ibn_sina": 0.026912588183717872, "hypatia": 0.7865011143684814, "einstein": 0.030495957668404457, "sagan": 0.034556447256743936, "darwin": 0.06455994289063166}, "confidence": 0.5261370879146867, "choice": "hypatia"}}, "usage": {"input_tokens": 342, "output_tokens": 1}, "latency_ms": 186.1258610006189, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 342, "compiled_prompt_sha256": "ebf830a50a0084bf727dee0ecca59607345f1d5b5ff40a355f83f10100363532", "compiled_ids_sha256": "735e6ae15336ca10f18bed8cf9f335970bb12077fe9c321d6e337aa2879ee706", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.25, 16.5, 19.875, 16.625, 16.75, 17.375], "runtime_probabilities": [0.05697395280003548, 0.026912588626146317, 0.7865011096000671, 0.030495958402752876, 0.034556448459625244, 0.0645599439740181], "decide_seconds": 0.18205066300288308, "probabilities_f64_minus_runtime_max_abs": 4.768414219036288e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:36Z"} {"id": "c-ibn_sina-01-dessert-02", "uid": "c057", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "prompt_sha256_by_order": ["fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e"], "request_sha256": "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "compiled_prompt_sha256_by_order": ["319fd600c27e869175120d52429454dc20b5548f94e0e370905b96ceb450ebfc", "fc07f5d1d7837dc5f780d230c04d7ba189bd12134f043747eb2c9bb060d658e3", "b1de02c44ffe4e7cb5658167df659bd0c9653ca47e976d09991fbeec1f0d4408", "cb3a577fa0c69484f0047139df1cc7529c904331ef0dd79451cd3e3aa328e91d", "6dba88eb5598bf7f49e6233f1684908940026c2b713555bb23dc984a684cec8b", "233a7140192e50df93982d534c857e749e0fe6edc924badc6e732211857a206a"], "compiled_tokens_by_order": [339, 339, 339, 339, 339, 339], "choice": "ibn_sina", "probs": {"darwin": 0.012670491923410142, "einstein": 0.01190282562512413, "hypatia": 0.005281846853512981, "ibn_sina": 0.9455585578784056, "sagan": 0.019624441789576702, "socrates": 0.004961835929970349}, "confidence": 0.8369257959962197, "seconds": 0.18574643216561526, "seconds_all_orders": 1.1144785929936916, "seconds_by_order": [0.18620589999773074, 0.18536558599953423, 0.18558064499666216, 0.18548780600031023, 0.1858753929991508, 0.1859632630003034], "argmax_stable": true, "spread": {"darwin": 0.10336085854537759, "einstein": 0.05829949037923095, "hypatia": 0.007970233107507612, "ibn_sina": 0.11827508855671087, "sagan": 0.018578170202203585, "socrates": 0.03805869309336442}, "spread_2dp": {"darwin": 0.11, "einstein": 0.060000000000000005, "hypatia": 0.01, "ibn_sina": 0.12, "sagan": 0.02, "socrates": 0.04}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.012670491923410142, "einstein": 0.01190282562512413, "hypatia": 0.005281846853512981, "ibn_sina": 0.9455585578784056, "sagan": 0.019624441789576702, "socrates": 0.004961835929970349}, {"ibn_sina": 0.8272834693216947, "hypatia": 0.007157404423151503, "sagan": 0.015152225282715362, "socrates": 0.04118801864672289, "darwin": 0.09880492033907969, "einstein": 0.010413961986635776}, {"hypatia": 0.00966061578241847, "darwin": 0.02795391864620779, "sagan": 0.005170955004070413, "ibn_sina": 0.9257066499856873, "einstein": 0.02795391864620779, "socrates": 0.003553941935408123}, {"ibn_sina": 0.8676752451963238, "einstein": 0.013174952685308883, "darwin": 0.09145229801521802, "sagan": 0.01492917725458639, "hypatia": 0.009639001295204504, "socrates": 0.003129325553358471}, {"einstein": 0.011488659486972327, "sagan": 0.0037298216490854784, "socrates": 0.007417636068395505, "ibn_sina": 0.8573621578485556, "darwin": 0.11603135046878774, "hypatia": 0.0039703744782033645}, {"socrates": 0.012710710611101895, "ibn_sina": 0.8371012184662571, "hypatia": 0.011940607585710977, "einstein": 0.06871345236586672, "sagan": 0.022307991851289063, "darwin": 0.04722601911977436}], "prompt_tokens": 339, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.012670491923410142, "einstein": 0.01190282562512413, "hypatia": 0.005281846853512981, "ibn_sina": 0.9455585578784056, "sagan": 0.019624441789576702, "socrates": 0.004961835929970349}, "confidence": 0.8369257959962197, "choice": "ibn_sina"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 186.17938099851017, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "319fd600c27e869175120d52429454dc20b5548f94e0e370905b96ceb450ebfc", "compiled_ids_sha256": "7cd4b48291d7d001e70cad1045450e5ebc69a94b89e94cfc8233e6b63d6d448e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 15.875, 15.0625, 20.25, 16.375, 15.0], "runtime_probabilities": [0.012670492753386497, 0.01190282590687275, 0.005281846970319748, 0.9455585479736328, 0.019624441862106323, 0.004961836151778698], "decide_seconds": 0.18194275299902074, "probabilities_f64_minus_runtime_max_abs": 9.904772779734117e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.8272834693216947, "hypatia": 0.007157404423151503, "sagan": 0.015152225282715362, "socrates": 0.04118801864672289, "darwin": 0.09880492033907969, "einstein": 0.010413961986635776}, "confidence": 0.6298051979910235, "choice": "ibn_sina"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 185.34183700103313, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "fc07f5d1d7837dc5f780d230c04d7ba189bd12134f043747eb2c9bb060d658e3", "compiled_ids_sha256": "81f8dcad8d99c88fc716915bfa3dcb041f1df8ebdd94e5ff0b44d693f0fe059a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.5, 14.75, 15.5, 16.5, 17.375, 15.125], "runtime_probabilities": [0.8272834420204163, 0.007157404441386461, 0.015152225270867348, 0.041188016533851624, 0.09880492091178894, 0.010413962416350842], "decide_seconds": 0.18136802799926954, "probabilities_f64_minus_runtime_max_abs": 2.7301278460178935e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.00966061578241847, "darwin": 0.02795391864620779, "sagan": 0.005170955004070413, "ibn_sina": 0.9257066499856873, "einstein": 0.02795391864620779, "socrates": 0.003553941935408123}, "confidence": 0.7971016676227818, "choice": "ibn_sina"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 185.5576550005935, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "b1de02c44ffe4e7cb5658167df659bd0c9653ca47e976d09991fbeec1f0d4408", "compiled_ids_sha256": "c2db1c0fc322cab49cd6dd4fde364dfba39de91c1e5ae1c48a6063304bbfc372", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.6875, 16.75, 15.0625, 20.25, 16.75, 14.6875], "runtime_probabilities": [0.009660616517066956, 0.027953920885920525, 0.005170955788344145, 0.9257067441940308, 0.027953920885920525, 0.003553942311555147], "decide_seconds": 0.18160491600065143, "probabilities_f64_minus_runtime_max_abs": 9.420834345963414e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.8676752451963238, "einstein": 0.013174952685308883, "darwin": 0.09145229801521802, "sagan": 0.01492917725458639, "hypatia": 0.009639001295204504, "socrates": 0.003129325553358471}, "confidence": 0.7072691316084827, "choice": "ibn_sina"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 185.46470599903842, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "cb3a577fa0c69484f0047139df1cc7529c904331ef0dd79451cd3e3aa328e91d", "compiled_ids_sha256": "51611cd39337ce5f4a25d3628163e7c9e812a7adbe100e8a34f3da1348be6203", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.625, 15.4375, 17.375, 15.5625, 15.125, 14.0], "runtime_probabilities": [0.867675244808197, 0.013174952939152718, 0.09145229309797287, 0.014929177239537239, 0.00963900238275528, 0.0031293255742639303], "decide_seconds": 0.18142300700128544, "probabilities_f64_minus_runtime_max_abs": 4.917245152724625e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.011488659486972327, "sagan": 0.0037298216490854784, "socrates": 0.007417636068395505, "ibn_sina": 0.8573621578485556, "darwin": 0.11603135046878774, "hypatia": 0.0039703744782033645}, "confidence": 0.7140476415139634, "choice": "ibn_sina"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 185.85176300257444, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "6dba88eb5598bf7f49e6233f1684908940026c2b713555bb23dc984a684cec8b", "compiled_ids_sha256": "47f5ea364455f61ec4c0c6af4a5f10bfe2973af14c53619a907bdc142c2a2e51", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.8125, 14.6875, 15.375, 20.125, 18.125, 14.75], "runtime_probabilities": [0.011488660238683224, 0.0037298216484487057, 0.00741763599216938, 0.8573621511459351, 0.11603134870529175, 0.003970374818891287], "decide_seconds": 0.1818561440013582, "probabilities_f64_minus_runtime_max_abs": 6.702620569676299e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.012710710611101895, "ibn_sina": 0.8371012184662571, "hypatia": 0.011940607585710977, "einstein": 0.06871345236586672, "sagan": 0.022307991851289063, "darwin": 0.04722601911977436}, "confidence": 0.6259493048506961, "choice": "ibn_sina"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 185.94009300068137, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 339, "compiled_prompt_sha256": "233a7140192e50df93982d534c857e749e0fe6edc924badc6e732211857a206a", "compiled_ids_sha256": "1ea83d5358ad7332ffb1aa8de7a81adc06c9282cc30823f1ace50ed669c3eb96", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 20.125, 15.875, 17.625, 16.5, 17.25], "runtime_probabilities": [0.012710710056126118, 0.8371012210845947, 0.011940606869757175, 0.06871344894170761, 0.022307991981506348, 0.04722602292895317], "decide_seconds": 0.18198377299995627, "probabilities_f64_minus_runtime_max_abs": 3.809178812652636e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:37Z"} {"id": "c-ibn_sina-01-dessert-03", "uid": "c058", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "prompt_sha256_by_order": ["a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c"], "request_sha256": "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "compiled_prompt_sha256_by_order": ["31e37b8f5f725ef736f2d61439d483950b0c0100444ccf92ee30ba189e9af0ef", "7064ccd748648c0f2c9753d94b5d60388cc828d9f69d0e10739045a782a9026c", "b18c386c42d75a2b1507d314bf26f7105eede4ceeefde8f9b570ab7db106cdd1", "aa2c4e11f21d36ddf82ad40b63631513ad43985be9fcfa74dc72ea6a452c492b", "6012266a2fa2c2dd0247aced474f41b0d142b78c9f2081c7fb744f049f787e5d", "a7a04181e9aebfc106fb6edbfdd1814d937a130ef09de73af59f3cc5ec826651"], "compiled_tokens_by_order": [351, 351, 351, 351, 351, 351], "choice": "hypatia", "probs": {"darwin": 0.01617218663666248, "einstein": 0.006741570383976906, "hypatia": 0.4726205716329198, "ibn_sina": 0.41708619056381263, "sagan": 0.08212914048845606, "socrates": 0.005250340294172206}, "confidence": 0.41276521227225127, "seconds": 0.18692096616662943, "seconds_all_orders": 1.1215257969997765, "seconds_by_order": [0.18700610399901052, 0.18671141600134433, 0.18702700399808236, 0.1870368130003044, 0.18677265600126702, 0.18697180399976787], "argmax_stable": false, "spread": {"darwin": 0.09295558718210808, "einstein": 0.060558462027202165, "hypatia": 0.6118790122825253, "ibn_sina": 0.47039841729021364, "sagan": 0.33177192937940236, "socrates": 0.03011500756698258}, "spread_2dp": {"darwin": 0.09000000000000001, "einstein": 0.060000000000000005, "hypatia": 0.61, "ibn_sina": 0.47, "sagan": 0.32999999999999996, "socrates": 0.03}, "choices_by_order": ["hypatia", "sagan", "ibn_sina", "ibn_sina", "ibn_sina", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.01617218663666248, "einstein": 0.006741570383976906, "hypatia": 0.4726205716329198, "ibn_sina": 0.41708619056381263, "sagan": 0.08212914048845606, "socrates": 0.005250340294172206}, {"ibn_sina": 0.32234647732718036, "hypatia": 0.08150195303648816, "sagan": 0.4139010698678584, "socrates": 0.033975068750533396, "darwin": 0.10465057920854144, "einstein": 0.043624851809398305}, {"hypatia": 0.2349001941633125, "darwin": 0.011694992026433355, "sagan": 0.18294045515801285, "ibn_sina": 0.4972837149460531, "einstein": 0.06730003241117907, "socrates": 0.005880611295009057}, {"ibn_sina": 0.3264193923009923, "einstein": 0.016251464601086108, "darwin": 0.07283401127967355, "sagan": 0.3264193923009923, "hypatia": 0.25421567833370495, "socrates": 0.003860061183550815}, {"einstein": 0.048787148370185794, "sagan": 0.28075065456575976, "socrates": 0.0203375092039507, "ibn_sina": 0.3604909762141559, "darwin": 0.0709848820223247, "hypatia": 0.2186488296236232}, {"socrates": 0.023726191906343203, "ibn_sina": 0.026885297655839455, "hypatia": 0.6933809653190134, "einstein": 0.05022834865988406, "sagan": 0.17531416300996427, "darwin": 0.030465033448955653}], "prompt_tokens": 351, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.01617218663666248, "einstein": 0.006741570383976906, "hypatia": 0.4726205716329198, "ibn_sina": 0.41708619056381263, "sagan": 0.08212914048845606, "socrates": 0.005250340294172206}, "confidence": 0.41276521227225127, "choice": "hypatia"}}, "usage": {"input_tokens": 351, "output_tokens": 1}, "latency_ms": 186.97893400167231, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 351, "compiled_prompt_sha256": "31e37b8f5f725ef736f2d61439d483950b0c0100444ccf92ee30ba189e9af0ef", "compiled_ids_sha256": "a7851da9192bd84a58d364630f143cf2398c247083217825c24b2c65fd8f0f9e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 15.25, 19.5, 19.375, 17.75, 15.0], "runtime_probabilities": [0.01617218740284443, 0.006741570308804512, 0.47262057662010193, 0.4170861840248108, 0.08212913572788239, 0.005250340793281794], "decide_seconds": 0.18265591800081893, "probabilities_f64_minus_runtime_max_abs": 6.539001840000225e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.32234647732718036, "hypatia": 0.08150195303648816, "sagan": 0.4139010698678584, "socrates": 0.033975068750533396, "darwin": 0.10465057920854144, "einstein": 0.043624851809398305}, "confidence": 0.2062864755655135, "choice": "sagan"}}, "usage": {"input_tokens": 351, "output_tokens": 1}, "latency_ms": 186.68778600113, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 351, "compiled_prompt_sha256": "7064ccd748648c0f2c9753d94b5d60388cc828d9f69d0e10739045a782a9026c", "compiled_ids_sha256": "2324df4a215a5745c4707c4d8f8edf1135993030acac8d65f6e7a78ccd5b039b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.875, 17.5, 19.125, 16.625, 17.75, 16.875], "runtime_probabilities": [0.3223464787006378, 0.08150194585323334, 0.41390106081962585, 0.03397506847977638, 0.1046505719423294, 0.04362484812736511], "decide_seconds": 0.1826180479984032, "probabilities_f64_minus_runtime_max_abs": 9.048232552100188e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.2349001941633125, "darwin": 0.011694992026433355, "sagan": 0.18294045515801285, "ibn_sina": 0.4972837149460531, "einstein": 0.06730003241117907, "socrates": 0.005880611295009057}, "confidence": 0.29551835458895703, "choice": "ibn_sina"}}, "usage": {"input_tokens": 351, "output_tokens": 1}, "latency_ms": 187.0026840006176, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 351, "compiled_prompt_sha256": "b18c386c42d75a2b1507d314bf26f7105eede4ceeefde8f9b570ab7db106cdd1", "compiled_ids_sha256": "abf1efbfa2153e30dab711df02d4d549b99c25d5d3d9bb83ebf6cff664d7c286", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.75, 15.75, 18.5, 19.5, 17.5, 15.0625], "runtime_probabilities": [0.23490022122859955, 0.011694992892444134, 0.18294048309326172, 0.4972837567329407, 0.06730003654956818, 0.0058806114830076694], "decide_seconds": 0.18287787600274896, "probabilities_f64_minus_runtime_max_abs": 4.178688756129034e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.3264193923009923, "einstein": 0.016251464601086108, "darwin": 0.07283401127967355, "sagan": 0.3264193923009923, "hypatia": 0.25421567833370495, "socrates": 0.003860061183550815}, "confidence": 0.24194032731214343, "choice": "ibn_sina"}}, "usage": {"input_tokens": 351, "output_tokens": 1}, "latency_ms": 187.01305500144372, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 351, "compiled_prompt_sha256": "aa2c4e11f21d36ddf82ad40b63631513ad43985be9fcfa74dc72ea6a452c492b", "compiled_ids_sha256": "12e049f0c86c059791799a19d3065dccdb225bd542b964eb7bfd3b24a9e6646e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.0, 16.0, 17.5, 19.0, 18.75, 14.5625], "runtime_probabilities": [0.3264194130897522, 0.016251465305685997, 0.07283401489257812, 0.3264194130897522, 0.25421568751335144, 0.003860061289742589], "decide_seconds": 0.18285228600143455, "probabilities_f64_minus_runtime_max_abs": 2.078875988020812e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.048787148370185794, "sagan": 0.28075065456575976, "socrates": 0.0203375092039507, "ibn_sina": 0.3604909762141559, "darwin": 0.0709848820223247, "hypatia": 0.2186488296236232}, "confidence": 0.17891015698016477, "choice": "ibn_sina"}}, "usage": {"input_tokens": 351, "output_tokens": 1}, "latency_ms": 186.7489159994875, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 351, "compiled_prompt_sha256": "6012266a2fa2c2dd0247aced474f41b0d142b78c9f2081c7fb744f049f787e5d", "compiled_ids_sha256": "6a623731292bdcae5d1ab665f7d81dcf414d34e8b7dcda00141decf2cd544432", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 18.75, 16.125, 19.0, 17.375, 18.5], "runtime_probabilities": [0.04878715053200722, 0.28075066208839417, 0.020337510854005814, 0.36049097776412964, 0.07098487764596939, 0.21864883601665497], "decide_seconds": 0.18257994899977348, "probabilities_f64_minus_runtime_max_abs": 7.522634404288908e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.023726191906343203, "ibn_sina": 0.026885297655839455, "hypatia": 0.6933809653190134, "einstein": 0.05022834865988406, "sagan": 0.17531416300996427, "darwin": 0.030465033448955653}, "confidence": 0.44091913646694914, "choice": "hypatia"}}, "usage": {"input_tokens": 351, "output_tokens": 1}, "latency_ms": 186.9477439977345, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 351, "compiled_prompt_sha256": "a7a04181e9aebfc106fb6edbfdd1814d937a130ef09de73af59f3cc5ec826651", "compiled_ids_sha256": "d994e60a3d465bf24d7e166b1d1657d7fbe5a5c9d247dfe92152ecb34d2e61be", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 16.75, 20.0, 17.375, 18.625, 16.875], "runtime_probabilities": [0.023726193234324455, 0.026885299012064934, 0.6933810114860535, 0.05022835358977318, 0.17531417310237885, 0.030465036630630493], "decide_seconds": 0.18284737700014375, "probabilities_f64_minus_runtime_max_abs": 4.616704007975869e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:38Z"} {"id": "c-ibn_sina-01-dessert-03", "uid": "c059", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "prompt_sha256_by_order": ["5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de"], "request_sha256": "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "compiled_prompt_sha256_by_order": ["ba26f55ee8938b9cb0c3d4aabab1559ed19fcddd55502d02f38ac6dfdc9310ac", "f73da195f3d4ff8ad2b788d1861c94edc583a4558283eea237bcf2009ffd6d7a", "786941d7d59b0639240e0e5dfed6a5f5228060470d42e0be5ef919db367cebb4", "3313925cad3e57296a8ab4edabadbdda056a31d1195be93d79db72de90d06c2f", "307b482d6dcdd9895913213838827eb440c78b88ba5d7775b66b7bda68288dfe", "84a25c35093b3dbfe4158880bb0ee9aa7a46ad8c80652e1837ee591bd4a4f445"], "compiled_tokens_by_order": [393, 393, 393, 393, 393, 393], "choice": "ibn_sina", "probs": {"darwin": 0.006104885283994854, "einstein": 0.001861884816576152, "hypatia": 0.008344388832677732, "ibn_sina": 0.9644802130417103, "sagan": 0.017665071297401375, "socrates": 0.0015435567276396037}, "confidence": 0.8889686333763394, "seconds": 0.21158798433316406, "seconds_all_orders": 1.2695279059989844, "seconds_by_order": [0.21138120400064508, 0.21188352999888593, 0.2115833930001827, 0.21182589100135374, 0.21120726499793818, 0.21164662299997872], "argmax_stable": true, "spread": {"darwin": 0.013205504772418471, "einstein": 0.012086832159638546, "hypatia": 0.04999873141895211, "ibn_sina": 0.08293141282335514, "sagan": 0.01555258047206052, "socrates": 0.005818169753947411}, "spread_2dp": {"darwin": 0.02, "einstein": 0.01, "hypatia": 0.05, "ibn_sina": 0.07999999999999996, "sagan": 0.02, "socrates": 0.01}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.006104885283994854, "einstein": 0.001861884816576152, "hypatia": 0.008344388832677732, "ibn_sina": 0.9644802130417103, "sagan": 0.017665071297401375, "socrates": 0.0015435567276396037}, {"ibn_sina": 0.989721139693268, "hypatia": 0.0011588443127330628, "sagan": 0.0023046371061445814, "socrates": 0.0010886334851485438, "darwin": 0.004878916791994265, "einstein": 0.0008478286107114385}, {"hypatia": 0.00950724682520744, "darwin": 0.0024038084445014082, "sagan": 0.008931231858991815, "ibn_sina": 0.9697655405528514, "einstein": 0.008390115875352794, "socrates": 0.0010020564430951106}, {"ibn_sina": 0.9854300225551118, "einstein": 0.0014815341178461283, "darwin": 0.0033386886730094835, "sagan": 0.005171062174463676, "hypatia": 0.004286961114551645, "socrates": 0.000291731365017249}, {"einstein": 0.004472143112340711, "sagan": 0.002112490825340854, "socrates": 0.001363926588473764, "ibn_sina": 0.9657140126336455, "darwin": 0.01560931321691988, "hypatia": 0.010728113623279256}, {"socrates": 0.00610990111896466, "ibn_sina": 0.9067897268699129, "hypatia": 0.05115757573168517, "einstein": 0.012934660770349984, "sagan": 0.014656890842866248, "darwin": 0.008351244666221131}], "prompt_tokens": 393, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.006104885283994854, "einstein": 0.001861884816576152, "hypatia": 0.008344388832677732, "ibn_sina": 0.9644802130417103, "sagan": 0.017665071297401375, "socrates": 0.0015435567276396037}, "confidence": 0.8889686333763394, "choice": "ibn_sina"}}, "usage": {"input_tokens": 393, "output_tokens": 1}, "latency_ms": 211.3519449994783, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 393, "compiled_prompt_sha256": "ba26f55ee8938b9cb0c3d4aabab1559ed19fcddd55502d02f38ac6dfdc9310ac", "compiled_ids_sha256": "8f0bad5ac64183cef3d192f03fe419ee35d679774c7b48af85a668fb1670c955", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5625, 14.375, 15.875, 20.625, 16.625, 14.1875], "runtime_probabilities": [0.006104885134845972, 0.0018618848407641053, 0.00834438856691122, 0.9644802212715149, 0.017665071412920952, 0.0015435568057000637], "decide_seconds": 0.20662362200164353, "probabilities_f64_minus_runtime_max_abs": 8.229804615567105e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.989721139693268, "hypatia": 0.0011588443127330628, "sagan": 0.0023046371061445814, "socrates": 0.0010886334851485438, "darwin": 0.004878916791994265, "einstein": 0.0008478286107114385}, "confidence": 0.9601232944255774, "choice": "ibn_sina"}}, "usage": {"input_tokens": 393, "output_tokens": 1}, "latency_ms": 211.85821000108263, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 393, "compiled_prompt_sha256": "f73da195f3d4ff8ad2b788d1861c94edc583a4558283eea237bcf2009ffd6d7a", "compiled_ids_sha256": "7b0b3941c0fe9f2cafbd86699092f7ac1408b7805ac1da098aac84e343d31302", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.375, 13.625, 14.3125, 13.5625, 15.0625, 13.3125], "runtime_probabilities": [0.9897212386131287, 0.0011588443303480744, 0.002304637338966131, 0.001088633551262319, 0.004878917243331671, 0.0008478286908939481], "decide_seconds": 0.20725719699839829, "probabilities_f64_minus_runtime_max_abs": 9.891986063603753e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.00950724682520744, "darwin": 0.0024038084445014082, "sagan": 0.008931231858991815, "ibn_sina": 0.9697655405528514, "einstein": 0.008390115875352794, "socrates": 0.0010020564430951106}, "confidence": 0.9008225403295329, "choice": "ibn_sina"}}, "usage": {"input_tokens": 393, "output_tokens": 1}, "latency_ms": 211.55786300005275, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 393, "compiled_prompt_sha256": "786941d7d59b0639240e0e5dfed6a5f5228060470d42e0be5ef919db367cebb4", "compiled_ids_sha256": "70b73c50c8d1f27c82fc4e875de53af42b6930960831645f28dfc49eba7593ba", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 14.625, 15.9375, 20.625, 15.875, 13.75], "runtime_probabilities": [0.009507247246801853, 0.0024038085248321295, 0.008931231684982777, 0.9697656035423279, 0.00839011650532484, 0.0010020564077422023], "decide_seconds": 0.20704630800173618, "probabilities_f64_minus_runtime_max_abs": 6.298947652361164e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.9854300225551118, "einstein": 0.0014815341178461283, "darwin": 0.0033386886730094835, "sagan": 0.005171062174463676, "hypatia": 0.004286961114551645, "socrates": 0.000291731365017249}, "confidence": 0.9463517999936781, "choice": "ibn_sina"}}, "usage": {"input_tokens": 393, "output_tokens": 1}, "latency_ms": 211.80111100329668, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 393, "compiled_prompt_sha256": "3313925cad3e57296a8ab4edabadbdda056a31d1195be93d79db72de90d06c2f", "compiled_ids_sha256": "752b3313009471cb89ab4fc593cb304d6d06b617638281f8ff8de33b64971599", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.375, 13.875, 14.6875, 15.125, 14.9375, 12.25], "runtime_probabilities": [0.9854300022125244, 0.0014815340982750058, 0.003338688751682639, 0.005171061959117651, 0.004286961164325476, 0.0002917313831858337], "decide_seconds": 0.20729068699802156, "probabilities_f64_minus_runtime_max_abs": 2.0342587392363498e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.004472143112340711, "sagan": 0.002112490825340854, "socrates": 0.001363926588473764, "ibn_sina": 0.9657140126336455, "darwin": 0.01560931321691988, "hypatia": 0.010728113623279256}, "confidence": 0.892016748232896, "choice": "ibn_sina"}}, "usage": {"input_tokens": 393, "output_tokens": 1}, "latency_ms": 211.18194600057905, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 393, "compiled_prompt_sha256": "307b482d6dcdd9895913213838827eb440c78b88ba5d7775b66b7bda68288dfe", "compiled_ids_sha256": "e44ae453b8a6f00b59c6a14c2f07974bdb8015a2f67b29c167404aba28aa9c8c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.25, 14.5, 14.0625, 20.625, 16.5, 16.125], "runtime_probabilities": [0.004472143482416868, 0.002112490823492408, 0.0013639265671372414, 0.9657139778137207, 0.015609312802553177, 0.010728114284574986], "decide_seconds": 0.2066147510013252, "probabilities_f64_minus_runtime_max_abs": 3.481992483234109e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.00610990111896466, "ibn_sina": 0.9067897268699129, "hypatia": 0.05115757573168517, "einstein": 0.012934660770349984, "sagan": 0.014656890842866248, "darwin": 0.008351244666221131}, "confidence": 0.7599841524557355, "choice": "ibn_sina"}}, "usage": {"input_tokens": 393, "output_tokens": 1}, "latency_ms": 211.6214529996796, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 393, "compiled_prompt_sha256": "84a25c35093b3dbfe4158880bb0ee9aa7a46ad8c80652e1837ee591bd4a4f445", "compiled_ids_sha256": "609a4b39d35c3b9de92d40c6499ecf122e9e9f6d389fecf4cc5b20e58e2b5e98", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.625, 20.625, 17.75, 16.375, 16.5, 15.9375], "runtime_probabilities": [0.006109901703894138, 0.9067897796630859, 0.05115758255124092, 0.012934661470353603, 0.01465689204633236, 0.00835124496370554], "decide_seconds": 0.2069771389978996, "probabilities_f64_minus_runtime_max_abs": 5.279317305539877e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:39Z"} {"id": "c-ibn_sina-01-main-02", "uid": "c060", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "prompt_sha256_by_order": ["1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3"], "request_sha256": "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "compiled_prompt_sha256_by_order": ["ca53ab37565565807844480c303f0ac73c53fed29285230b4137aee8eb216a3b", "39e98b0948b727153d60ea5539b099ddb23e1bf077549d483ccddb71bd1d69b1", "ee11c68407ba0ef55c0c9352ac37e4063aca3c082e92b4c62f5e4899f6ee8bae", "75b52dfb0ca8e8c86961c20509cb18423a814d5ba413449941bfde6d18d78b60", "5aae23cbb9249a5311d01d124a3ea27e3e8e8c45865e504e191d918851f97541", "dd62f54b1437e983be0f8cf41ea0eaab16198c22cff47431dd9de6e3abfa4ab2"], "compiled_tokens_by_order": [317, 317, 317, 317, 317, 317], "choice": "sagan", "probs": {"darwin": 0.027403508456313706, "einstein": 0.031052243215875568, "hypatia": 0.15769630309464852, "ibn_sina": 0.3338430762711283, "sagan": 0.42866299511735295, "socrates": 0.02134187384468098}, "confidence": 0.269355775252889, "seconds": 0.17770225299985518, "seconds_all_orders": 1.066213517999131, "seconds_by_order": [0.17807330299910973, 0.1775492779997876, 0.17776563599909423, 0.17777604599905317, 0.1770787810019101, 0.17797047400017618], "argmax_stable": false, "spread": {"darwin": 0.019290685417954825, "einstein": 0.17482258144314697, "hypatia": 0.18806894339811292, "ibn_sina": 0.3370697655463395, "sagan": 0.19370539648617663, "socrates": 0.04854671338150771}, "spread_2dp": {"darwin": 0.019999999999999997, "einstein": 0.18, "hypatia": 0.18, "ibn_sina": 0.33999999999999997, "sagan": 0.19999999999999998, "socrates": 0.049999999999999996}, "choices_by_order": ["sagan", "ibn_sina", "sagan", "ibn_sina", "ibn_sina", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.027403508456313706, "einstein": 0.031052243215875568, "hypatia": 0.15769630309464852, "ibn_sina": 0.3338430762711283, "sagan": 0.42866299511735295, "socrates": 0.02134187384468098}, {"ibn_sina": 0.3630260306544169, "hypatia": 0.13354981328783377, "sagan": 0.3203693476101033, "socrates": 0.02629751745771468, "darwin": 0.02320747770209753, "einstein": 0.13354981328783377}, {"hypatia": 0.20683762417458454, "darwin": 0.01166898026006323, "sagan": 0.3410175905577165, "ibn_sina": 0.30094696739404664, "einstein": 0.1254533606440045, "socrates": 0.01407547696958457}, {"ibn_sina": 0.41692093727943047, "einstein": 0.04394314349285765, "darwin": 0.030201651399014335, "sagan": 0.32469835243208456, "hypatia": 0.17379850396056012, "socrates": 0.0104374114360528}, {"einstein": 0.12576373988471498, "sagan": 0.23495759863117632, "socrates": 0.04082950845778642, "ibn_sina": 0.49740524020547444, "darwin": 0.024764348700643743, "hypatia": 0.07627956412020422}, {"socrates": 0.05898412481756051, "ibn_sina": 0.16033547465913492, "hypatia": 0.2643485075183171, "einstein": 0.20587482465902254, "sagan": 0.29954610236490536, "darwin": 0.010910965981059511}], "prompt_tokens": 317, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.027403508456313706, "einstein": 0.031052243215875568, "hypatia": 0.15769630309464852, "ibn_sina": 0.3338430762711283, "sagan": 0.42866299511735295, "socrates": 0.02134187384468098}, "confidence": 0.269355775252889, "choice": "sagan"}}, "usage": {"input_tokens": 317, "output_tokens": 1}, "latency_ms": 178.0437429988524, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 317, "compiled_prompt_sha256": "ca53ab37565565807844480c303f0ac73c53fed29285230b4137aee8eb216a3b", "compiled_ids_sha256": "46e60b1b41b20e85a4c648d7194e57c8c2da9d2e1ec5c2d9640a9a1b77e6ad6e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 16.5, 18.125, 18.875, 19.125, 16.125], "runtime_probabilities": [0.027403507381677628, 0.03105224296450615, 0.15769632160663605, 0.33384308218955994, 0.4286629855632782, 0.021341873332858086], "decide_seconds": 0.17406327499702456, "probabilities_f64_minus_runtime_max_abs": 1.8511987526226292e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.3630260306544169, "hypatia": 0.13354981328783377, "sagan": 0.3203693476101033, "socrates": 0.02629751745771468, "darwin": 0.02320747770209753, "einstein": 0.13354981328783377}, "confidence": 0.188910060479912, "choice": "ibn_sina"}}, "usage": {"input_tokens": 317, "output_tokens": 1}, "latency_ms": 177.52418800228043, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 317, "compiled_prompt_sha256": "39e98b0948b727153d60ea5539b099ddb23e1bf077549d483ccddb71bd1d69b1", "compiled_ids_sha256": "ea70e9e66ceee653cfc500254e77e108320a0811b68c1bc20786f61d5376abfd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.125, 18.125, 19.0, 16.5, 16.375, 18.125], "runtime_probabilities": [0.3630260229110718, 0.13354982435703278, 0.32036933302879333, 0.026297518983483315, 0.02320747636258602, 0.13354982435703278], "decide_seconds": 0.1736777969999821, "probabilities_f64_minus_runtime_max_abs": 1.4581309992767189e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.20683762417458454, "darwin": 0.01166898026006323, "sagan": 0.3410175905577165, "ibn_sina": 0.30094696739404664, "einstein": 0.1254533606440045, "socrates": 0.01407547696958457}, "confidence": 0.20382161936932897, "choice": "sagan"}}, "usage": {"input_tokens": 317, "output_tokens": 1}, "latency_ms": 177.7401760009525, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 317, "compiled_prompt_sha256": "ee11c68407ba0ef55c0c9352ac37e4063aca3c082e92b4c62f5e4899f6ee8bae", "compiled_ids_sha256": "cedb33e032eafb968adad0ac88f71c2d1e7bc3d7be8a80e51d0c9a209b8eaa4c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.625, 15.75, 19.125, 19.0, 18.125, 15.9375], "runtime_probabilities": [0.20683760941028595, 0.011668979190289974, 0.34101757407188416, 0.3009469509124756, 0.12545336782932281, 0.014075476676225662], "decide_seconds": 0.1737938469996152, "probabilities_f64_minus_runtime_max_abs": 1.6485832365908948e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.41692093727943047, "einstein": 0.04394314349285765, "darwin": 0.030201651399014335, "sagan": 0.32469835243208456, "hypatia": 0.17379850396056012, "socrates": 0.0104374114360528}, "confidence": 0.26064500381527433, "choice": "ibn_sina"}}, "usage": {"input_tokens": 317, "output_tokens": 1}, "latency_ms": 177.75118599820416, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 317, "compiled_prompt_sha256": "75b52dfb0ca8e8c86961c20509cb18423a814d5ba413449941bfde6d18d78b60", "compiled_ids_sha256": "f2b09fd1c84bd7c66b5e3210182bf1687c414d239b9c42cc230a6c4673b94ee6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.125, 16.875, 16.5, 18.875, 18.25, 15.4375], "runtime_probabilities": [0.4169209599494934, 0.043943144381046295, 0.030201653018593788, 0.3246983587741852, 0.17379850149154663, 0.010437412187457085], "decide_seconds": 0.17391814599977806, "probabilities_f64_minus_runtime_max_abs": 2.2670062937990565e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.12576373988471498, "sagan": 0.23495759863117632, "socrates": 0.04082950845778642, "ibn_sina": 0.49740524020547444, "darwin": 0.024764348700643743, "hypatia": 0.07627956412020422}, "confidence": 0.23712720527721753, "choice": "ibn_sina"}}, "usage": {"input_tokens": 317, "output_tokens": 1}, "latency_ms": 177.05459100034204, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 317, "compiled_prompt_sha256": "5aae23cbb9249a5311d01d124a3ea27e3e8e8c45865e504e191d918851f97541", "compiled_ids_sha256": "85e5b10bfeb721d4715841608bd82f139bbe59df9105f7ed4f8ec89dfccce547", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 18.375, 16.625, 19.125, 16.125, 17.25], "runtime_probabilities": [0.12576374411582947, 0.23495760560035706, 0.04082950949668884, 0.4974052309989929, 0.024764347821474075, 0.07627956569194794], "decide_seconds": 0.1731772320017626, "probabilities_f64_minus_runtime_max_abs": 9.206481521584209e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.05898412481756051, "ibn_sina": 0.16033547465913492, "hypatia": 0.2643485075183171, "einstein": 0.20587482465902254, "sagan": 0.29954610236490536, "darwin": 0.010910965981059511}, "confidence": 0.13608101938600892, "choice": "sagan"}}, "usage": {"input_tokens": 317, "output_tokens": 1}, "latency_ms": 177.94565499934833, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 317, "compiled_prompt_sha256": "dd62f54b1437e983be0f8cf41ea0eaab16198c22cff47431dd9de6e3abfa4ab2", "compiled_ids_sha256": "db1b58832167ad7caebac5f82379980ec19eb64bd8a248074fe1a57eaa539002", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 18.125, 18.625, 18.375, 18.75, 15.4375], "runtime_probabilities": [0.058984119445085526, 0.1603354811668396, 0.26434850692749023, 0.20587483048439026, 0.29954609274864197, 0.010910965502262115], "decide_seconds": 0.1739696049990016, "probabilities_f64_minus_runtime_max_abs": 9.616263396683422e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:40Z"} {"id": "c-ibn_sina-01-main-02", "uid": "c061", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "prompt_sha256_by_order": ["0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c"], "request_sha256": "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "compiled_prompt_sha256_by_order": ["45e5e8ec908f7291f32e80dc6bdc7bd7e15ac15096efb3d82a2feb0b9cc1ed9e", "43ebe18dbb17a2f7624a1dd7f6cca575d005e584f01fb9b1e2f20ea142d0b939", "066fbd33d05df3931f35e871f7ba2a9e5622956e46f7f969d1fe1edfe24c21cd", "fb91358cbe69e57c48bca479dd8f25387d2ac5b7735a3ec39700029a75f69438", "5f495cfa45655914b778c03fdb85692182eca55025caf29f24329f6cbe089f84", "b7f7f1c571d02fc4949e129890bf2703743efafaeab498becfd9f3ee28d5edd1"], "compiled_tokens_by_order": [352, 352, 352, 352, 352, 352], "choice": "ibn_sina", "probs": {"darwin": 0.05538360194284894, "einstein": 0.05538360194284894, "hypatia": 0.11724708623308097, "ibn_sina": 0.6747103961907623, "sagan": 0.07111395256233462, "socrates": 0.026161361128124117}, "confidence": 0.3745795852321142, "seconds": 0.187076091666313, "seconds_all_orders": 1.122456549997878, "seconds_by_order": [0.18739783100318164, 0.18694762399900355, 0.18696384400027455, 0.18703969399939524, 0.18686906499715406, 0.187238491998869], "argmax_stable": true, "spread": {"darwin": 0.038469872390527515, "einstein": 0.18069144362185635, "hypatia": 0.08483106630384332, "ibn_sina": 0.2797008877168914, "sagan": 0.06537713436836128, "socrates": 0.03472286766191154}, "spread_2dp": {"darwin": 0.039999999999999994, "einstein": 0.18, "hypatia": 0.07999999999999999, "ibn_sina": 0.2799999999999999, "sagan": 0.060000000000000005, "socrates": 0.04}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.05538360194284894, "einstein": 0.05538360194284894, "hypatia": 0.11724708623308097, "ibn_sina": 0.6747103961907623, "sagan": 0.07111395256233462, "socrates": 0.026161361128124117}, {"ibn_sina": 0.806342520587405, "hypatia": 0.04014543020040087, "sagan": 0.0454907321292865, "socrates": 0.031265292476810626, "darwin": 0.031265292476810626, "einstein": 0.0454907321292865}, {"hypatia": 0.08255305070919693, "darwin": 0.02087267997376209, "sagan": 0.014345569165601715, "ibn_sina": 0.6912082309202776, "einstein": 0.17476480972279684, "socrates": 0.016255659508364745}, {"ibn_sina": 0.6892165113020183, "einstein": 0.09327531176840873, "darwin": 0.04992667662271263, "sagan": 0.07264288584646614, "hypatia": 0.08231517372323316, "socrates": 0.012623440737160947}, {"einstein": 0.02535996450227125, "sagan": 0.007265751478104856, "socrates": 0.011253416031634537, "ibn_sina": 0.839806686149106, "darwin": 0.04737865116199293, "hypatia": 0.06893553067689037}, {"socrates": 0.045976283693546074, "ibn_sina": 0.5601057984322146, "hypatia": 0.1249764965042442, "einstein": 0.20605140812412762, "sagan": 0.045976283693546074, "darwin": 0.016913729552321424}], "prompt_tokens": 352, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.05538360194284894, "einstein": 0.05538360194284894, "hypatia": 0.11724708623308097, "ibn_sina": 0.6747103961907623, "sagan": 0.07111395256233462, "socrates": 0.026161361128124117}, "confidence": 0.3745795852321142, "choice": "ibn_sina"}}, "usage": {"input_tokens": 352, "output_tokens": 1}, "latency_ms": 187.36916100169765, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 352, "compiled_prompt_sha256": "45e5e8ec908f7291f32e80dc6bdc7bd7e15ac15096efb3d82a2feb0b9cc1ed9e", "compiled_ids_sha256": "8cb23d0f97d0241542c05f0a7bceae8684c28142b6f39ba5f8ef5db574533fac", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 17.125, 17.875, 19.625, 17.375, 16.375], "runtime_probabilities": [0.055383604019880295, 0.055383604019880295, 0.1172470822930336, 0.6747104525566101, 0.0711139515042305, 0.02616136148571968], "decide_seconds": 0.18304135400103405, "probabilities_f64_minus_runtime_max_abs": 5.636584776702591e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.806342520587405, "hypatia": 0.04014543020040087, "sagan": 0.0454907321292865, "socrates": 0.031265292476810626, "darwin": 0.031265292476810626, "einstein": 0.0454907321292865}, "confidence": 0.5532441276686988, "choice": "ibn_sina"}}, "usage": {"input_tokens": 352, "output_tokens": 1}, "latency_ms": 186.92374400052358, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 352, "compiled_prompt_sha256": "43ebe18dbb17a2f7624a1dd7f6cca575d005e584f01fb9b1e2f20ea142d0b939", "compiled_ids_sha256": "a8ee1a7e9630195985a411143fd0c6e80e1295719119b12ea1fe94ee441b41bc", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.875, 16.875, 17.0, 16.625, 16.625, 17.0], "runtime_probabilities": [0.8063424825668335, 0.040145426988601685, 0.04549073055386543, 0.03126528859138489, 0.03126528859138489, 0.04549073055386543], "decide_seconds": 0.1827097480017983, "probabilities_f64_minus_runtime_max_abs": 3.802057146895521e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.08255305070919693, "darwin": 0.02087267997376209, "sagan": 0.014345569165601715, "ibn_sina": 0.6912082309202776, "einstein": 0.17476480972279684, "socrates": 0.016255659508364745}, "confidence": 0.4560412824104644, "choice": "ibn_sina"}}, "usage": {"input_tokens": 352, "output_tokens": 1}, "latency_ms": 186.94053400031407, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 352, "compiled_prompt_sha256": "066fbd33d05df3931f35e871f7ba2a9e5622956e46f7f969d1fe1edfe24c21cd", "compiled_ids_sha256": "0b5fcc8216b374c3e0a4862bbe036d81d4ce232c3505647362254d74cc3e8ea7", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 16.375, 16.0, 19.875, 18.5, 16.125], "runtime_probabilities": [0.08255305141210556, 0.020872680470347404, 0.014345568604767323, 0.6912082433700562, 0.17476481199264526, 0.016255659982562065], "decide_seconds": 0.18282134699984454, "probabilities_f64_minus_runtime_max_abs": 1.2449778541423484e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.6892165113020183, "einstein": 0.09327531176840873, "darwin": 0.04992667662271263, "sagan": 0.07264288584646614, "hypatia": 0.08231517372323316, "socrates": 0.012623440737160947}, "confidence": 0.39798394490032873, "choice": "ibn_sina"}}, "usage": {"input_tokens": 352, "output_tokens": 1}, "latency_ms": 187.0161039987579, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 352, "compiled_prompt_sha256": "fb91358cbe69e57c48bca479dd8f25387d2ac5b7735a3ec39700029a75f69438", "compiled_ids_sha256": "2f5846c01b3a8524e3d9fc66159650b093a31d98ac143b8b91d700661e92f2dd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.625, 17.625, 17.0, 17.375, 17.5, 15.625], "runtime_probabilities": [0.6892164945602417, 0.09327530860900879, 0.04992667958140373, 0.07264288514852524, 0.08231517672538757, 0.012623441405594349], "decide_seconds": 0.18284313600088353, "probabilities_f64_minus_runtime_max_abs": 1.6741776631690186e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.02535996450227125, "sagan": 0.007265751478104856, "socrates": 0.011253416031634537, "ibn_sina": 0.839806686149106, "darwin": 0.04737865116199293, "hypatia": 0.06893553067689037}, "confidence": 0.6344718355822073, "choice": "ibn_sina"}}, "usage": {"input_tokens": 352, "output_tokens": 1}, "latency_ms": 186.84361499981605, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 352, "compiled_prompt_sha256": "5f495cfa45655914b778c03fdb85692182eca55025caf29f24329f6cbe089f84", "compiled_ids_sha256": "5faa349e6b2578dcfc39544f5f485b1ba45f81024969b1902bafc7b968d831b4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 15.25, 15.6875, 20.0, 17.125, 17.5], "runtime_probabilities": [0.02535996399819851, 0.007265751250088215, 0.011253416538238525, 0.8398066759109497, 0.047378651797771454, 0.06893552839756012], "decide_seconds": 0.1826839370005473, "probabilities_f64_minus_runtime_max_abs": 1.0238156322728287e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.045976283693546074, "ibn_sina": 0.5601057984322146, "hypatia": 0.1249764965042442, "einstein": 0.20605140812412762, "sagan": 0.045976283693546074, "darwin": 0.016913729552321424}, "confidence": 0.2955386960223455, "choice": "ibn_sina"}}, "usage": {"input_tokens": 352, "output_tokens": 1}, "latency_ms": 187.21413299863343, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 352, "compiled_prompt_sha256": "b7f7f1c571d02fc4949e129890bf2703743efafaeab498becfd9f3ee28d5edd1", "compiled_ids_sha256": "70dfa8a1bfc1a14d9b1dae26fb2a6a808bd5e8ba25f6e508b3b99bda4d4530df", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 19.5, 18.0, 18.5, 17.0, 16.0], "runtime_probabilities": [0.04597628116607666, 0.5601058006286621, 0.1249764934182167, 0.20605142414569855, 0.04597628116607666, 0.016913728788495064], "decide_seconds": 0.18289955599902896, "probabilities_f64_minus_runtime_max_abs": 1.6021570931057383e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:41Z"} {"id": "c-ibn_sina-01-main-02", "uid": "c062", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "prompt_sha256_by_order": ["c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de"], "request_sha256": "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "compiled_prompt_sha256_by_order": ["2c19043fdf98431a6068bfeda9e535e1c58513f3ad737c128fbdbff9a34cc6d6", "f8a96c99bb0bcb4ffb9a654fd9d37a52f317b95761ec49f9d334f4d4ea2c2559", "6a1aa8de9a9f2ff697666631a887e22076082309700ab0e3c41d489443820ec7", "3c84328cb34dea419c7c25ce54449f742b7fbab8e99d65bb1e2c8b8b33532292", "c86433d742b975ae4f4c0e95443299abc7f3ec1b30de25134bfbc69e36d8cbc4", "248a411ff2de31f890240ce1d14cd50f09d8b969ac7249c4ad6ef1227d5c37a9"], "compiled_tokens_by_order": [365, 365, 365, 365, 365, 365], "choice": "ibn_sina", "probs": {"darwin": 0.002399906926500843, "einstein": 0.001755809415445316, "hypatia": 0.06588626223560634, "ibn_sina": 0.9095317948740986, "sagan": 0.01887673017768961, "socrates": 0.0015494963706593604}, "confidence": 0.7901362359985314, "seconds": 0.1941609501664061, "seconds_all_orders": 1.1649657009984367, "seconds_by_order": [0.19440279699847451, 0.19407449899881613, 0.19395888000144623, 0.19439302699902328, 0.19445220600027824, 0.19368429200039827], "argmax_stable": false, "spread": {"darwin": 0.0030129936938920136, "einstein": 0.01360802514663342, "hypatia": 0.5279873499798123, "ibn_sina": 0.5422121138593262, "sagan": 0.01487983644839068, "socrates": 0.002698095311837945}, "spread_2dp": {"darwin": 0.0, "einstein": 0.01, "hypatia": 0.52, "ibn_sina": 0.55, "sagan": 0.01, "socrates": 0.0}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.002399906926500843, "einstein": 0.001755809415445316, "hypatia": 0.06588626223560634, "ibn_sina": 0.9095317948740986, "sagan": 0.01887673017768961, "socrates": 0.0015494963706593604}, {"ibn_sina": 0.9609712098802423, "hypatia": 0.01553265295028251, "sagan": 0.01994432117675258, "socrates": 0.001742715362553341, "darwin": 0.0009328083144622913, "einstein": 0.0008762923157068969}, {"hypatia": 0.0287495574016103, "darwin": 0.0011147413137600594, "sagan": 0.0105763711108307, "ibn_sina": 0.9520545869666929, "einstein": 0.0068286184226623685, "socrates": 0.0006761247844438165}, {"ibn_sina": 0.9651199563496456, "einstein": 0.0021111913324959037, "darwin": 0.0014509991683008612, "sagan": 0.0050644847283619005, "hypatia": 0.025719575658020865, "socrates": 0.0005337927631747485}, {"einstein": 0.003706737949923682, "sagan": 0.00574110804593877, "socrates": 0.001061998203356842, "ibn_sina": 0.9655059177562564, "darwin": 0.003945802008354305, "hypatia": 0.0200384360361699}, {"socrates": 0.0032318880750126936, "ibn_sina": 0.42329380389693017, "hypatia": 0.5435200029300948, "einstein": 0.014484317462340317, "sagan": 0.014484317462340317, "darwin": 0.000985670173281627}], "prompt_tokens": 365, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.002399906926500843, "einstein": 0.001755809415445316, "hypatia": 0.06588626223560634, "ibn_sina": 0.9095317948740986, "sagan": 0.01887673017768961, "socrates": 0.0015494963706593604}, "confidence": 0.7901362359985314, "choice": "ibn_sina"}}, "usage": {"input_tokens": 365, "output_tokens": 1}, "latency_ms": 194.37515699974028, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 365, "compiled_prompt_sha256": "2c19043fdf98431a6068bfeda9e535e1c58513f3ad737c128fbdbff9a34cc6d6", "compiled_ids_sha256": "3d9a463164b741f3095477f08090e3b5559a38fc08b00bb779a1cc9e5a019438", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.5625, 14.25, 17.875, 20.5, 16.625, 14.125], "runtime_probabilities": [0.0023999069817364216, 0.001755809411406517, 0.06588626652956009, 0.9095317721366882, 0.018876729533076286, 0.0015494965482503176], "decide_seconds": 0.19002834999992047, "probabilities_f64_minus_runtime_max_abs": 2.2737410398931956e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.9609712098802423, "hypatia": 0.01553265295028251, "sagan": 0.01994432117675258, "socrates": 0.001742715362553341, "darwin": 0.0009328083144622913, "einstein": 0.0008762923157068969}, "confidence": 0.8857136992569539, "choice": "ibn_sina"}}, "usage": {"input_tokens": 365, "output_tokens": 1}, "latency_ms": 194.04987999951118, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 365, "compiled_prompt_sha256": "f8a96c99bb0bcb4ffb9a654fd9d37a52f317b95761ec49f9d334f4d4ea2c2559", "compiled_ids_sha256": "ad58a7c79fd06d1f6f83b57b67c4e720c704b32d0d1e5314b865e5c53d7f6575", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.25, 16.125, 16.375, 13.9375, 13.3125, 13.25], "runtime_probabilities": [0.9609712362289429, 0.015532652847468853, 0.019944321364164352, 0.0017427153652533889, 0.0009328082669526339, 0.0008762923534959555], "decide_seconds": 0.1897675820000586, "probabilities_f64_minus_runtime_max_abs": 2.6348700554734705e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0287495574016103, "darwin": 0.0011147413137600594, "sagan": 0.0105763711108307, "ibn_sina": 0.9520545869666929, "einstein": 0.0068286184226623685, "socrates": 0.0006761247844438165}, "confidence": 0.8641041145940489, "choice": "ibn_sina"}}, "usage": {"input_tokens": 365, "output_tokens": 1}, "latency_ms": 193.93530000161263, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 365, "compiled_prompt_sha256": "6a1aa8de9a9f2ff697666631a887e22076082309700ab0e3c41d489443820ec7", "compiled_ids_sha256": "7c1ce254745608e98c221fb7b287465928144449e665abbbcaec4193030f7365", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 14.125, 16.375, 20.875, 15.9375, 13.625], "runtime_probabilities": [0.028749560937285423, 0.0011147414334118366, 0.010576372966170311, 0.9520546793937683, 0.006828618701547384, 0.0006761248223483562], "decide_seconds": 0.18961395399674075, "probabilities_f64_minus_runtime_max_abs": 9.242707543766215e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.9651199563496456, "einstein": 0.0021111913324959037, "darwin": 0.0014509991683008612, "sagan": 0.0050644847283619005, "hypatia": 0.025719575658020865, "socrates": 0.0005337927631747485}, "confidence": 0.8985963935425065, "choice": "ibn_sina"}}, "usage": {"input_tokens": 365, "output_tokens": 1}, "latency_ms": 194.36922699969728, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 365, "compiled_prompt_sha256": "3c84328cb34dea419c7c25ce54449f742b7fbab8e99d65bb1e2c8b8b33532292", "compiled_ids_sha256": "694b9761221e04b9caad2dd622fb93a70d59f49fd89be0956251e67b53042bd4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.375, 14.25, 13.875, 15.125, 16.75, 12.875], "runtime_probabilities": [0.9651198983192444, 0.002111191162839532, 0.0014509990578517318, 0.005064484663307667, 0.025719573721289635, 0.0005337927141226828], "decide_seconds": 0.19011888000022736, "probabilities_f64_minus_runtime_max_abs": 5.803040126206582e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.003706737949923682, "sagan": 0.00574110804593877, "socrates": 0.001061998203356842, "ibn_sina": 0.9655059177562564, "darwin": 0.003945802008354305, "hypatia": 0.0200384360361699}, "confidence": 0.8929930756670298, "choice": "ibn_sina"}}, "usage": {"input_tokens": 365, "output_tokens": 1}, "latency_ms": 194.42737600184046, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 365, "compiled_prompt_sha256": "c86433d742b975ae4f4c0e95443299abc7f3ec1b30de25134bfbc69e36d8cbc4", "compiled_ids_sha256": "607de14276f46f3b59e16d3b9f5efdf6ae84efd14a524bad73f359d95a73868c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.9375, 15.375, 13.6875, 20.5, 15.0, 16.625], "runtime_probabilities": [0.00370673811994493, 0.005741107743233442, 0.0010619981912896037, 0.9655058979988098, 0.0039458018727600574, 0.020038435235619545], "decide_seconds": 0.19002404099956038, "probabilities_f64_minus_runtime_max_abs": 1.975744656768086e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0032318880750126936, "ibn_sina": 0.42329380389693017, "hypatia": 0.5435200029300948, "einstein": 0.014484317462340317, "sagan": 0.014484317462340317, "darwin": 0.000985670173281627}, "confidence": 0.5293402905962002, "choice": "hypatia"}}, "usage": {"input_tokens": 365, "output_tokens": 1}, "latency_ms": 193.66033199912636, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 365, "compiled_prompt_sha256": "248a411ff2de31f890240ce1d14cd50f09d8b969ac7249c4ad6ef1227d5c37a9", "compiled_ids_sha256": "7291562e49e886c24323a5132473b11430f918e9dae4dd1eb1321f1528c7426c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.875, 19.75, 20.0, 16.375, 16.375, 13.6875], "runtime_probabilities": [0.003231887938454747, 0.423293799161911, 0.5435200333595276, 0.014484317973256111, 0.014484317973256111, 0.0009856701362878084], "decide_seconds": 0.18940218599891523, "probabilities_f64_minus_runtime_max_abs": 3.042943275133325e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:43Z"} {"id": "c-ibn_sina-01-main-03", "uid": "c063", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "prompt_sha256_by_order": ["9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf"], "request_sha256": "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "compiled_prompt_sha256_by_order": ["7b9355b2e2fdcf6d623298737a222d53b12dc9d6c71de4bf927676f9c9b5c339", "839704d4054c24973da64c8cfcf8e745b6d22abc74db8054480aee813f8306a9", "6a8aba645314a1fa85e7ee3fb5056567e43519e1f258201c67d2c8e5ef9f4acb", "b7eee888397e052b348ea35ecf9cfeb4181bf95fe56f71b61bb5c95fc6063f01", "da835f51ed0a9ecd870385d5d74d76bcb254737df1ae2bf3e36c28195a5a1b74", "a3ac943ebad6c98f5b9a2b2a20b87f37c802031c79fe24ffe29857101c8d108c"], "compiled_tokens_by_order": [390, 390, 390, 390, 390, 390], "choice": "ibn_sina", "probs": {"darwin": 0.0015794514278882493, "einstein": 0.002604075165197046, "hypatia": 0.00403327051577787, "ibn_sina": 0.9869087558499959, "sagan": 0.001085540032758686, "socrates": 0.0037889070083821767}, "confidence": 0.9500707931448142, "seconds": 0.21155850616620833, "seconds_all_orders": 1.26935103699725, "seconds_by_order": [0.21137960399937583, 0.21189909000167972, 0.21155511299730279, 0.2118051010002091, 0.2116199919983046, 0.211092137000378], "argmax_stable": true, "spread": {"darwin": 0.0033386638522759867, "einstein": 0.004483488083613414, "hypatia": 0.0049394995590927755, "ibn_sina": 0.016288285328447505, "sagan": 0.000814854303589549, "socrates": 0.005751019143289875}, "spread_2dp": {"darwin": 0.0, "einstein": 0.0, "hypatia": 0.01, "ibn_sina": 0.020000000000000018, "sagan": 0.0, "socrates": 0.01}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0015794514278882493, "einstein": 0.002604075165197046, "hypatia": 0.00403327051577787, "ibn_sina": 0.9869087558499959, "sagan": 0.001085540032758686, "socrates": 0.0037889070083821767}, {"ibn_sina": 0.9929862544723672, "hypatia": 0.001092224914227731, "sagan": 0.0005846258678535669, "socrates": 0.004319833334456023, "darwin": 0.0006624678977811202, "einstein": 0.0003545935133142948}, {"hypatia": 0.0058481961665979324, "darwin": 0.001016264109941499, "sagan": 0.0005110096860371177, "ibn_sina": 0.9835153705292202, "einstein": 0.004554579754101657, "socrates": 0.004554579754101657}, {"ibn_sina": 0.9965068307416055, "einstein": 0.0009673025035266964, "darwin": 0.0004292378839746658, "sagan": 0.00033429080016338255, "hypatia": 0.0009086966075051568, "socrates": 0.0008536414632246343}, {"einstein": 0.004838081596927709, "sagan": 0.0011491451037529314, "socrates": 0.005482264677369714, "ibn_sina": 0.9814374452742146, "darwin": 0.0037679017362506526, "hypatia": 0.00332516161148432}, {"socrates": 0.00660466060651451, "ibn_sina": 0.980218545413158, "hypatia": 0.005475455958499552, "einstein": 0.004539312424910502, "sagan": 0.001147717918745835, "darwin": 0.002014307678171659}], "prompt_tokens": 390, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0015794514278882493, "einstein": 0.002604075165197046, "hypatia": 0.00403327051577787, "ibn_sina": 0.9869087558499959, "sagan": 0.001085540032758686, "socrates": 0.0037889070083821767}, "confidence": 0.9500707931448142, "choice": "ibn_sina"}}, "usage": {"input_tokens": 390, "output_tokens": 1}, "latency_ms": 211.34883500053547, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 390, "compiled_prompt_sha256": "7b9355b2e2fdcf6d623298737a222d53b12dc9d6c71de4bf927676f9c9b5c339", "compiled_ids_sha256": "45337ce0418c787e52b602dcd036092f44cf3be9bf63859bfa2bb7606fe58827", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.6875, 15.1875, 15.625, 21.125, 14.3125, 15.5625], "runtime_probabilities": [0.001579451491124928, 0.0026040752418339252, 0.004033270757645369, 0.9869087338447571, 0.001085540046915412, 0.003788907080888748], "decide_seconds": 0.20670518099723267, "probabilities_f64_minus_runtime_max_abs": 2.200523885331762e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.9929862544723672, "hypatia": 0.001092224914227731, "sagan": 0.0005846258678535669, "socrates": 0.004319833334456023, "darwin": 0.0006624678977811202, "einstein": 0.0003545935133142948}, "confidence": 0.9721082144637381, "choice": "ibn_sina"}}, "usage": {"input_tokens": 390, "output_tokens": 1}, "latency_ms": 211.87184999871533, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 390, "compiled_prompt_sha256": "839704d4054c24973da64c8cfcf8e745b6d22abc74db8054480aee813f8306a9", "compiled_ids_sha256": "b5aa16ab2d17d63b1810c22bd88461f8d6f7e39e05c2a0a0b4eadacac0e39e4b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.625, 13.8125, 13.1875, 15.1875, 13.3125, 12.6875], "runtime_probabilities": [0.992986261844635, 0.0010922248475253582, 0.0005846258718520403, 0.004319833591580391, 0.0006624679081141949, 0.0003545935032889247], "decide_seconds": 0.20732144500288996, "probabilities_f64_minus_runtime_max_abs": 7.372267796235121e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0058481961665979324, "darwin": 0.001016264109941499, "sagan": 0.0005110096860371177, "ibn_sina": 0.9835153705292202, "einstein": 0.004554579754101657, "socrates": 0.004554579754101657}, "confidence": 0.9406130795334287, "choice": "ibn_sina"}}, "usage": {"input_tokens": 390, "output_tokens": 1}, "latency_ms": 211.52805400197394, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 390, "compiled_prompt_sha256": "6a8aba645314a1fa85e7ee3fb5056567e43519e1f258201c67d2c8e5ef9f4acb", "compiled_ids_sha256": "848f0fa51b134d4a7cf5177855ec47ac62d13963b35d46a3885bcdd595f8984d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 14.25, 13.5625, 21.125, 15.75, 15.75], "runtime_probabilities": [0.005848195869475603, 0.0010162640828639269, 0.0005110096535645425, 0.9835153818130493, 0.004554579965770245, 0.004554579965770245], "decide_seconds": 0.20699524800147628, "probabilities_f64_minus_runtime_max_abs": 1.1283829093322595e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.9965068307416055, "einstein": 0.0009673025035266964, "darwin": 0.0004292378839746658, "sagan": 0.00033429080016338255, "hypatia": 0.0009086966075051568, "socrates": 0.0008536414632246343}, "confidence": 0.9840377096827517, "choice": "ibn_sina"}}, "usage": {"input_tokens": 390, "output_tokens": 1}, "latency_ms": 211.7778409992752, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 390, "compiled_prompt_sha256": "b7eee888397e052b348ea35ecf9cfeb4181bf95fe56f71b61bb5c95fc6063f01", "compiled_ids_sha256": "6ea46cdfe48e2dfc197a3a837b76cb45f1f557976675947c87105bc0f7a710c1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.625, 13.6875, 12.875, 12.625, 13.625, 13.5625], "runtime_probabilities": [0.9965068697929382, 0.000967302534263581, 0.00042923790169879794, 0.00033429081668145955, 0.0009086966747418046, 0.0008536414825357497], "decide_seconds": 0.207194526999956, "probabilities_f64_minus_runtime_max_abs": 3.9051332723083476e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.004838081596927709, "sagan": 0.0011491451037529314, "socrates": 0.005482264677369714, "ibn_sina": 0.9814374452742146, "darwin": 0.0037679017362506526, "hypatia": 0.00332516161148432}, "confidence": 0.9327442666337311, "choice": "ibn_sina"}}, "usage": {"input_tokens": 390, "output_tokens": 1}, "latency_ms": 211.58967199880863, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 390, "compiled_prompt_sha256": "da835f51ed0a9ecd870385d5d74d76bcb254737df1ae2bf3e36c28195a5a1b74", "compiled_ids_sha256": "91546ae5eebe1f1aed50cb1c711035ba6e2d70be59095a20191a73e8344d8d5e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5625, 14.125, 15.6875, 20.875, 15.3125, 15.1875], "runtime_probabilities": [0.004838081542402506, 0.0011491450713947415, 0.005482264794409275, 0.9814374446868896, 0.003767901798710227, 0.0033251617569476366], "decide_seconds": 0.2068510190001689, "probabilities_f64_minus_runtime_max_abs": 5.873249664389846e-10}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.00660466060651451, "ibn_sina": 0.980218545413158, "hypatia": 0.005475455958499552, "einstein": 0.004539312424910502, "sagan": 0.001147717918745835, "darwin": 0.002014307678171659}, "confidence": 0.9296689155929949, "choice": "ibn_sina"}}, "usage": {"input_tokens": 390, "output_tokens": 1}, "latency_ms": 211.06540700202459, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 390, "compiled_prompt_sha256": "a3ac943ebad6c98f5b9a2b2a20b87f37c802031c79fe24ffe29857101c8d108c", "compiled_ids_sha256": "1774314ec5ff3f00a168eac60e6d655dd13b8a4c0f459b9de303e2d58363da99", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 21.0, 15.8125, 15.625, 14.25, 14.8125], "runtime_probabilities": [0.006604661233723164, 0.9802185893058777, 0.005475456826388836, 0.00453931326046586, 0.0011477179359644651, 0.0020143077708780766], "decide_seconds": 0.206500991997018, "probabilities_f64_minus_runtime_max_abs": 4.3892719681082326e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:44Z"} {"id": "c-ibn_sina-01-main-03", "uid": "c064", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "prompt_sha256_by_order": ["eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032"], "request_sha256": "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "compiled_prompt_sha256_by_order": ["0dc80c8adce05b741ce8a4c3f17b9cc1f78bc491ba12ae7dfad74cd0ca49ecbb", "92f29f32e14c0d979e049d00ae719f31c0d0da0122d96aa375125d516c967b4e", "1fdc75835ec0479b5bbe1e83c16b140a074ebf5531ff54502deb61b080c6ec67", "2fb060e55afe946d016fe92b383c97bff1d9af578c8faffb60ecf93e3304d0b5", "46b4e4f2c74a7952af24aff3e6d1e10460bdd83b6948a8b801d2554d33b0880c", "141aa7fb13c0d156a5285ca3b26ef84e029202c3c89660e85c97a44d86fff8f5"], "compiled_tokens_by_order": [371, 371, 371, 371, 371, 371], "choice": "ibn_sina", "probs": {"darwin": 0.015965222301884067, "einstein": 0.01323563416648501, "hypatia": 0.2829903364943132, "ibn_sina": 0.4117485100191416, "sagan": 0.026322201600572324, "socrates": 0.2497380954176038}, "confidence": 0.28109576428568983, "seconds": 0.1948027499984164, "seconds_all_orders": 1.1688164999904984, "seconds_by_order": [0.19506327099952614, 0.1947212839986605, 0.1949401919991942, 0.19483363299877965, 0.19488213299700874, 0.19437598699732916], "argmax_stable": false, "spread": {"darwin": 0.012286262642353607, "einstein": 0.014950798430873553, "hypatia": 0.37599519986713337, "ibn_sina": 0.5727116996609796, "sagan": 0.01986673981859665, "socrates": 0.3235414063088112}, "spread_2dp": {"darwin": 0.01, "einstein": 0.01, "hypatia": 0.38, "ibn_sina": 0.5700000000000001, "sagan": 0.019999999999999997, "socrates": 0.33}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.015965222301884067, "einstein": 0.01323563416648501, "hypatia": 0.2829903364943132, "ibn_sina": 0.4117485100191416, "sagan": 0.026322201600572324, "socrates": 0.2497380954176038}, {"ibn_sina": 0.7850243341875209, "hypatia": 0.04428804235138311, "sagan": 0.016292660270803987, "socrates": 0.1364167742563256, "darwin": 0.012688736577218512, "einstein": 0.0052894523567478735}, {"hypatia": 0.21760320310721032, "darwin": 0.009560817491162341, "sagan": 0.011532547267033226, "ibn_sina": 0.5915068328208126, "einstein": 0.020240250787621428, "socrates": 0.14955634852616}, {"ibn_sina": 0.9000286178125492, "einstein": 0.007314996532077752, "darwin": 0.006455461781975674, "sagan": 0.006455461781975674, "hypatia": 0.06519785439821423, "socrates": 0.01454760769320755}, {"einstein": 0.016832460854542215, "sagan": 0.010867853546890358, "socrates": 0.33808901400201874, "ibn_sina": 0.4919166127496701, "darwin": 0.017918061309611967, "hypatia": 0.12437599753726658}, {"socrates": 0.22496140861247593, "ibn_sina": 0.32731691815156955, "hypatia": 0.4202832422185165, "einstein": 0.011922517872145186, "sagan": 0.009884114478034587, "darwin": 0.00563179866725836}], "prompt_tokens": 371, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.015965222301884067, "einstein": 0.01323563416648501, "hypatia": 0.2829903364943132, "ibn_sina": 0.4117485100191416, "sagan": 0.026322201600572324, "socrates": 0.2497380954176038}, "confidence": 0.28109576428568983, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 195.03357099893037, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "0dc80c8adce05b741ce8a4c3f17b9cc1f78bc491ba12ae7dfad74cd0ca49ecbb", "compiled_ids_sha256": "5948b85d5511809552d45e91339724803cfce0ca4a60973ed39e61cc07e354a8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.875, 15.6875, 18.75, 19.125, 16.375, 18.625], "runtime_probabilities": [0.01596522144973278, 0.013235634192824364, 0.28299033641815186, 0.4117484986782074, 0.026322199031710625, 0.24973808228969574], "decide_seconds": 0.19058374599990202, "probabilities_f64_minus_runtime_max_abs": 1.3127908066179472e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.7850243341875209, "hypatia": 0.04428804235138311, "sagan": 0.016292660270803987, "socrates": 0.1364167742563256, "darwin": 0.012688736577218512, "einstein": 0.0052894523567478735}, "confidence": 0.5814056456689184, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 194.694794001407, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "92f29f32e14c0d979e049d00ae719f31c0d0da0122d96aa375125d516c967b4e", "compiled_ids_sha256": "349e335094b5b2613ba976f92a3d52b0bda3da0dc84aae7ab7bc6348372f525f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.625, 16.75, 15.75, 17.875, 15.5, 14.625], "runtime_probabilities": [0.7850244045257568, 0.04428804665803909, 0.016292661428451538, 0.13641677796840668, 0.012688737362623215, 0.005289452616125345], "decide_seconds": 0.19035019800139708, "probabilities_f64_minus_runtime_max_abs": 7.033823590241184e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.21760320310721032, "darwin": 0.009560817491162341, "sagan": 0.011532547267033226, "ibn_sina": 0.5915068328208126, "einstein": 0.020240250787621428, "socrates": 0.14955634852616}, "confidence": 0.38525015850742417, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 194.91387200105237, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "1fdc75835ec0479b5bbe1e83c16b140a074ebf5531ff54502deb61b080c6ec67", "compiled_ids_sha256": "7bbfab7f9a1495424321586c307489982b246392ed3e26f8727e7a1dd8c34555", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.75, 15.625, 15.8125, 19.75, 16.375, 18.375], "runtime_probabilities": [0.21760320663452148, 0.009560816921293736, 0.011532546952366829, 0.5915067791938782, 0.02024024911224842, 0.14955633878707886], "decide_seconds": 0.19054726700051106, "probabilities_f64_minus_runtime_max_abs": 5.36269344486584e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.9000286178125492, "einstein": 0.007314996532077752, "darwin": 0.006455461781975674, "sagan": 0.006455461781975674, "hypatia": 0.06519785439821423, "socrates": 0.01454760769320755}, "confidence": 0.756980174231263, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 194.80636300068, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "2fb060e55afe946d016fe92b383c97bff1d9af578c8faffb60ecf93e3304d0b5", "compiled_ids_sha256": "256699a7033ee5bfb677c22b959892945d72206216e42a670785a4afa3eb83d5", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.0, 15.1875, 15.0625, 15.0625, 17.375, 15.875], "runtime_probabilities": [0.900028645992279, 0.007314997259527445, 0.006455461494624615, 0.006455461494624615, 0.06519785523414612, 0.014547607861459255], "decide_seconds": 0.19029734899959294, "probabilities_f64_minus_runtime_max_abs": 2.8179729882893412e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.016832460854542215, "sagan": 0.010867853546890358, "socrates": 0.33808901400201874, "ibn_sina": 0.4919166127496701, "darwin": 0.017918061309611967, "hypatia": 0.12437599753726658}, "confidence": 0.34988859547010076, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 194.85583300047438, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "46b4e4f2c74a7952af24aff3e6d1e10460bdd83b6948a8b801d2554d33b0880c", "compiled_ids_sha256": "5d706c18d28ff7f3ad2f27cce23ac56a9bd9f104f4fa89d73ea315c1e28ef8de", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.875, 15.4375, 18.875, 19.25, 15.9375, 17.875], "runtime_probabilities": [0.016832459717988968, 0.010867851786315441, 0.3380889892578125, 0.49191656708717346, 0.0179180596023798, 0.12437598407268524], "decide_seconds": 0.1905430160004471, "probabilities_f64_minus_runtime_max_abs": 4.566249661719368e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.22496140861247593, "ibn_sina": 0.32731691815156955, "hypatia": 0.4202832422185165, "einstein": 0.011922517872145186, "sagan": 0.009884114478034587, "darwin": 0.00563179866725836}, "confidence": 0.33412767163118307, "choice": "hypatia"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 194.35071699990658, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "141aa7fb13c0d156a5285ca3b26ef84e029202c3c89660e85c97a44d86fff8f5", "compiled_ids_sha256": "708e0f0b5dfb29dc2304483dcab07e249867406883d1f8fa3310ccc684659188", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.625, 19.0, 19.25, 15.6875, 15.5, 14.9375], "runtime_probabilities": [0.22496141493320465, 0.32731691002845764, 0.4202832281589508, 0.011922517791390419, 0.009884114377200603, 0.005631798878312111], "decide_seconds": 0.19002362099854508, "probabilities_f64_minus_runtime_max_abs": 1.4059565678348207e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:45Z"} {"id": "c-ibn_sina-01-main-03", "uid": "c065", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "prompt_sha256_by_order": ["c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086"], "request_sha256": "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "compiled_prompt_sha256_by_order": ["667abdaec59327957c3b1ce1c6570cf662d99312a7e39336fd34c6a5b9174833", "1457bcd521f631b20130859f96550f0e527fce3c361cf06075a56425819cbf27", "3c56b6d22318a0264796bbd2a4497b08d207ec80950a4da84c215ed194a553a0", "99f47c3be7c674e23eb09726b53e8cfc0e27222cd22465dcfd6d5b65babaa6e2", "295b7162c7889b2500fdcaebdf11e9fdfbd4accabd136238e467217b468723b1", "504ae730d37539aad2381411fd9a1e632d32a31cba4b065b59b4f7bc3fff7132"], "compiled_tokens_by_order": [395, 395, 395, 395, 395, 395], "choice": "ibn_sina", "probs": {"darwin": 0.01168218178675752, "einstein": 0.008546866775387608, "hypatia": 0.021825188897098072, "ibn_sina": 0.9280306467603671, "sagan": 0.024731179036637956, "socrates": 0.005183936743751833}, "confidence": 0.7967081151083615, "seconds": 0.21204360433330294, "seconds_all_orders": 1.2722616259998176, "seconds_by_order": [0.21201777000169386, 0.21229869699891424, 0.21191779100263375, 0.2121279289967788, 0.21220668699970702, 0.21169275200008997], "argmax_stable": true, "spread": {"darwin": 0.02821610827367764, "einstein": 0.03500640040747549, "hypatia": 0.06196861154759273, "ibn_sina": 0.17918335465231638, "sagan": 0.0791507119634948, "socrates": 0.010269613372238555}, "spread_2dp": {"darwin": 0.03, "einstein": 0.03, "hypatia": 0.06, "ibn_sina": 0.18000000000000005, "sagan": 0.08, "socrates": 0.01}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.01168218178675752, "einstein": 0.008546866775387608, "hypatia": 0.021825188897098072, "ibn_sina": 0.9280306467603671, "sagan": 0.024731179036637956, "socrates": 0.005183936743751833}, {"ibn_sina": 0.8908839239565397, "hypatia": 0.0345433184214309, "sagan": 0.02690236343649527, "socrates": 0.013527357857920662, "darwin": 0.02690236343649527, "einstein": 0.007240672891118201}, {"hypatia": 0.019194422219319864, "darwin": 0.00906680305559566, "sagan": 0.014948631055009482, "ibn_sina": 0.9248390546654524, "einstein": 0.027927719537667244, "socrates": 0.0040233694669554905}, {"ibn_sina": 0.7488472921080507, "einstein": 0.042247073298593694, "darwin": 0.0372829113292733, "sagan": 0.08943705487495973, "hypatia": 0.07892792390344044, "socrates": 0.003257744485682107}, {"einstein": 0.016959312355847717, "sagan": 0.010286342911464932, "socrates": 0.005172298037722115, "ibn_sina": 0.9259470804635307, "darwin": 0.024675653875586814, "hypatia": 0.016959312355847717}, {"socrates": 0.011201309217587654, "ibn_sina": 0.8898302070204291, "hypatia": 0.026870543942156946, "einstein": 0.03044831530111932, "sagan": 0.03044831530111932, "darwin": 0.011201309217587654}], "prompt_tokens": 395, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.01168218178675752, "einstein": 0.008546866775387608, "hypatia": 0.021825188897098072, "ibn_sina": 0.9280306467603671, "sagan": 0.024731179036637956, "socrates": 0.005183936743751833}, "confidence": 0.7967081151083615, "choice": "ibn_sina"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 211.989370000083, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "667abdaec59327957c3b1ce1c6570cf662d99312a7e39336fd34c6a5b9174833", "compiled_ids_sha256": "5cf614b764b9c0e5e91de93bb9bc44413b47819378f9850fb3926ac12e4ae96b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 15.8125, 16.75, 20.5, 16.875, 15.3125], "runtime_probabilities": [0.011682181619107723, 0.008546866476535797, 0.021825188770890236, 0.9280306100845337, 0.02473117783665657, 0.005183937028050423], "decide_seconds": 0.20723631599685177, "probabilities_f64_minus_runtime_max_abs": 3.6675833370303224e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.8908839239565397, "hypatia": 0.0345433184214309, "sagan": 0.02690236343649527, "socrates": 0.013527357857920662, "darwin": 0.02690236343649527, "einstein": 0.007240672891118201}, "confidence": 0.716694666274923, "choice": "ibn_sina"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 212.2728080030356, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "1457bcd521f631b20130859f96550f0e527fce3c361cf06075a56425819cbf27", "compiled_ids_sha256": "7d4bf7e48f3e8b271911fdf7932a4c21070a4290a0a62adbd27723fb12f0f704", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.875, 16.625, 16.375, 15.6875, 16.375, 15.0625], "runtime_probabilities": [0.8908839821815491, 0.034543320536613464, 0.026902364566922188, 0.013527357950806618, 0.026902364566922188, 0.007240673992782831], "decide_seconds": 0.2075623140008247, "probabilities_f64_minus_runtime_max_abs": 5.82250093694725e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.019194422219319864, "darwin": 0.00906680305559566, "sagan": 0.014948631055009482, "ibn_sina": 0.9248390546654524, "einstein": 0.027927719537667244, "socrates": 0.0040233694669554905}, "confidence": 0.7902982508477823, "choice": "ibn_sina"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 211.89205100017716, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "3c56b6d22318a0264796bbd2a4497b08d207ec80950a4da84c215ed194a553a0", "compiled_ids_sha256": "bfbfd824b862da1bc19553b2be0dded328783033e6d2c26d9518916cd345f071", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 15.875, 16.375, 20.5, 17.0, 15.0625], "runtime_probabilities": [0.01919442228972912, 0.009066804312169552, 0.014948631636798382, 0.9248391389846802, 0.027927720919251442, 0.004023369867354631], "decide_seconds": 0.20721921700169332, "probabilities_f64_minus_runtime_max_abs": 8.43192278221494e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.7488472921080507, "einstein": 0.042247073298593694, "darwin": 0.0372829113292733, "sagan": 0.08943705487495973, "hypatia": 0.07892792390344044, "socrates": 0.003257744485682107}, "confidence": 0.4932998318335361, "choice": "ibn_sina"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 212.10221899673343, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "99f47c3be7c674e23eb09726b53e8cfc0e27222cd22465dcfd6d5b65babaa6e2", "compiled_ids_sha256": "0c282b5357c92272a6afe1f967802a734487fae6388a7a4b826c8f1f78b42064", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.5, 16.625, 16.5, 17.375, 17.25, 14.0625], "runtime_probabilities": [0.748847246170044, 0.04224707558751106, 0.03728291019797325, 0.08943705260753632, 0.0789279192686081, 0.0032577444799244404], "decide_seconds": 0.20733140599986655, "probabilities_f64_minus_runtime_max_abs": 4.593800673191595e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.016959312355847717, "sagan": 0.010286342911464932, "socrates": 0.005172298037722115, "ibn_sina": 0.9259470804635307, "darwin": 0.024675653875586814, "hypatia": 0.016959312355847717}, "confidence": 0.7906070629133044, "choice": "ibn_sina"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 212.18126799794845, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "295b7162c7889b2500fdcaebdf11e9fdfbd4accabd136238e467217b468723b1", "compiled_ids_sha256": "766600d72a832559ac0695e171dde43c726438684c7c5d244cb26104ae981f62", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 16.0, 15.3125, 20.5, 16.875, 16.5], "runtime_probabilities": [0.016959311440587044, 0.010286343283951283, 0.005172298289835453, 0.9259470701217651, 0.024675652384757996, 0.016959311440587044], "decide_seconds": 0.20752184400043916, "probabilities_f64_minus_runtime_max_abs": 1.0341765555033078e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.011201309217587654, "ibn_sina": 0.8898302070204291, "hypatia": 0.026870543942156946, "einstein": 0.03044831530111932, "sagan": 0.03044831530111932, "darwin": 0.011201309217587654}, "confidence": 0.712958625500767, "choice": "ibn_sina"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 211.6665120011021, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "504ae730d37539aad2381411fd9a1e632d32a31cba4b065b59b4f7bc3fff7132", "compiled_ids_sha256": "3576abe73dfbb8680d649f45c0bb1da7d0d2142b03e8f68de99b13f53e1b767b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 20.5, 17.0, 17.125, 17.125, 16.125], "runtime_probabilities": [0.011201309971511364, 0.8898302912712097, 0.026870546862483025, 0.030448317527770996, 0.030448317527770996, 0.011201309971511364], "decide_seconds": 0.20694080799876247, "probabilities_f64_minus_runtime_max_abs": 8.425078057427982e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:46Z"} {"id": "c-ibn_sina-01-starter-01", "uid": "c066", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "prompt_sha256_by_order": ["43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423"], "request_sha256": "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "compiled_prompt_sha256_by_order": ["0cc652ed158d4456d057ae4ea6f240eebcd0b64c04d9666f8ff066084455a03d", "051c8a009024080fc05ff9992a9126e6e5556581499cdf44658c8923a02ebecc", "6d83fc1bf8b4240715f520976bd2d427b922117c4f560bbf0823743856b3302e", "467fe3acfa3d4644b06eab8263530d26f099342c4ce4b607700c9b598b3b629c", "a77dd33c62eb50ff5d0c72ef241a7c36c6e40315ebbcdcddd425295a9513da25", "bd644fac9a14c031262ff3cac101ea851c45ce1bcebbbb073935e6c94ab052ef"], "compiled_tokens_by_order": [293, 293, 293, 293, 293, 293], "choice": "ibn_sina", "probs": {"darwin": 0.012603548346877986, "einstein": 0.004935620448730488, "hypatia": 0.4442903939829127, "ibn_sina": 0.5034469726541884, "sagan": 0.012603548346877986, "socrates": 0.022119916220412483}, "confidence": 0.48278976233978177, "seconds": 0.17485806383350186, "seconds_all_orders": 1.049148383001011, "seconds_by_order": [0.17547888399712974, 0.17535019500064664, 0.17462234100094065, 0.17442971199852764, 0.17433873300251435, 0.1749285180012521], "argmax_stable": false, "spread": {"darwin": 0.007518271716612661, "einstein": 0.007122788475013476, "hypatia": 0.5923574974516564, "ibn_sina": 0.626190213388167, "sagan": 0.011367289089276148, "socrates": 0.04162281283820085}, "spread_2dp": {"darwin": 0.0, "einstein": 0.01, "hypatia": 0.6000000000000001, "ibn_sina": 0.63, "sagan": 0.01, "socrates": 0.04}, "choices_by_order": ["ibn_sina", "ibn_sina", "hypatia", "ibn_sina", "ibn_sina", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.012603548346877986, "einstein": 0.004935620448730488, "hypatia": 0.4442903939829127, "ibn_sina": 0.5034469726541884, "sagan": 0.012603548346877986, "socrates": 0.022119916220412483}, {"ibn_sina": 0.5772697445280437, "hypatia": 0.3501317989807377, "sagan": 0.014451665278652213, "socrates": 0.047385186185202884, "darwin": 0.009330696391590931, "einstein": 0.001430908635772521}, {"hypatia": 0.9027907730233674, "darwin": 0.005368190163814406, "sagan": 0.004450385957731756, "ibn_sina": 0.07410557936129164, "einstein": 0.003255971921521221, "socrates": 0.010029099572273683}, {"ibn_sina": 0.6660398004917659, "einstein": 0.005413248794981922, "darwin": 0.005085276630265325, "sagan": 0.0030843761893760645, "hypatia": 0.31461492454660867, "socrates": 0.005762373347002036}, {"einstein": 0.008553697110785997, "sagan": 0.003795680084159591, "socrates": 0.02804649307544903, "ibn_sina": 0.5633288722328588, "darwin": 0.009105363177693396, "hypatia": 0.3871698943190532}, {"socrates": 0.031034889641354927, "ibn_sina": 0.03984958710359884, "hypatia": 0.9069724219982651, "einstein": 0.007371432408151771, "sagan": 0.006924819895864371, "darwin": 0.007846848952765094}], "prompt_tokens": 293, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.012603548346877986, "einstein": 0.004935620448730488, "hypatia": 0.4442903939829127, "ibn_sina": 0.5034469726541884, "sagan": 0.012603548346877986, "socrates": 0.022119916220412483}, "confidence": 0.48278976233978177, "choice": "ibn_sina"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 175.45292399881873, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 293, "compiled_prompt_sha256": "0cc652ed158d4456d057ae4ea6f240eebcd0b64c04d9666f8ff066084455a03d", "compiled_ids_sha256": "d87679564b3455f5e970ac7dffd32d0e50699c17f522bec77e89352be1475266", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.8125, 14.875, 19.375, 19.5, 15.8125, 16.375], "runtime_probabilities": [0.012603548355400562, 0.004935620818287134, 0.4442903697490692, 0.5034469366073608, 0.012603548355400562, 0.0221199169754982], "decide_seconds": 0.1717616529967927, "probabilities_f64_minus_runtime_max_abs": 3.604682752289534e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.5772697445280437, "hypatia": 0.3501317989807377, "sagan": 0.014451665278652213, "socrates": 0.047385186185202884, "darwin": 0.009330696391590931, "einstein": 0.001430908635772521}, "confidence": 0.47351215485905207, "choice": "ibn_sina"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 175.3246650005167, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 293, "compiled_prompt_sha256": "051c8a009024080fc05ff9992a9126e6e5556581499cdf44658c8923a02ebecc", "compiled_ids_sha256": "42010f20150d52a892158a49a35b8dc42661bbd9cef267f488ea8e85c38b6c78", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.375, 18.875, 15.6875, 16.875, 15.25, 13.375], "runtime_probabilities": [0.5772697329521179, 0.3501318097114563, 0.014451665803790092, 0.047385185956954956, 0.009330696426331997, 0.0014309086836874485], "decide_seconds": 0.17143563499848824, "probabilities_f64_minus_runtime_max_abs": 1.1575925773499307e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.9027907730233674, "darwin": 0.005368190163814406, "sagan": 0.004450385957731756, "ibn_sina": 0.07410557936129164, "einstein": 0.003255971921521221, "socrates": 0.010029099572273683}, "confidence": 0.7755676194691964, "choice": "hypatia"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 174.5991909992881, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 293, "compiled_prompt_sha256": "6d83fc1bf8b4240715f520976bd2d427b922117c4f560bbf0823743856b3302e", "compiled_ids_sha256": "fa864540899bcf7e47fb7c7d78cb1c292d4d5b4501554ccc6e6a0b7a78281c3f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.375, 15.25, 15.0625, 17.875, 14.75, 15.875], "runtime_probabilities": [0.9027908444404602, 0.005368190351873636, 0.004450386390089989, 0.07410558313131332, 0.003255972173064947, 0.010029100812971592], "decide_seconds": 0.1708358590003627, "probabilities_f64_minus_runtime_max_abs": 7.141709279512298e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.6660398004917659, "einstein": 0.005413248794981922, "darwin": 0.005085276630265325, "sagan": 0.0030843761893760645, "hypatia": 0.31461492454660867, "socrates": 0.005762373347002036}, "confidence": 0.5885838280116441, "choice": "ibn_sina"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 174.4075120004709, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 293, "compiled_prompt_sha256": "467fe3acfa3d4644b06eab8263530d26f099342c4ce4b607700c9b598b3b629c", "compiled_ids_sha256": "e5a64b699cf87fc6cc342b8d82c5b11e0deb42f8f8c463dc91054d882f080894", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.875, 15.0625, 15.0, 14.5, 19.125, 15.125], "runtime_probabilities": [0.666039764881134, 0.005413249135017395, 0.005085276439785957, 0.0030843764543533325, 0.3146149218082428, 0.005762373097240925], "decide_seconds": 0.1708684200020798, "probabilities_f64_minus_runtime_max_abs": 3.561063188772806e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.008553697110785997, "sagan": 0.003795680084159591, "socrates": 0.02804649307544903, "ibn_sina": 0.5633288722328588, "darwin": 0.009105363177693396, "hypatia": 0.3871698943190532}, "confidence": 0.500169239211675, "choice": "ibn_sina"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 174.31607300022733, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 293, "compiled_prompt_sha256": "a77dd33c62eb50ff5d0c72ef241a7c36c6e40315ebbcdcddd425295a9513da25", "compiled_ids_sha256": "9664907577861e4b904f89b63d1f5468ba3f85330b6520d9a18880b5d9189f8d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.3125, 14.5, 16.5, 19.5, 15.375, 19.125], "runtime_probabilities": [0.008553696796298027, 0.0037956801243126392, 0.028046492487192154, 0.5633288621902466, 0.009105362929403782, 0.38716989755630493], "decide_seconds": 0.17068334199939272, "probabilities_f64_minus_runtime_max_abs": 1.0042612186289546e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.031034889641354927, "ibn_sina": 0.03984958710359884, "hypatia": 0.9069724219982651, "einstein": 0.007371432408151771, "sagan": 0.006924819895864371, "darwin": 0.007846848952765094}, "confidence": 0.7581025786504606, "choice": "hypatia"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 174.9067679993459, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 293, "compiled_prompt_sha256": "bd644fac9a14c031262ff3cac101ea851c45ce1bcebbbb073935e6c94ab052ef", "compiled_ids_sha256": "6c7024adcc488c3914596d4009087d1e4d32391d116887aeb732f953b654898e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 17.0, 20.125, 15.3125, 15.25, 15.375], "runtime_probabilities": [0.031034888699650764, 0.03984958678483963, 0.9069724082946777, 0.007371433079242706, 0.00692481966689229, 0.007846849039196968], "decide_seconds": 0.17132268699788256, "probabilities_f64_minus_runtime_max_abs": 1.3703587375957227e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:47Z"} {"id": "c-ibn_sina-01-starter-02", "uid": "c067", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "prompt_sha256_by_order": ["131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b"], "request_sha256": "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "compiled_prompt_sha256_by_order": ["ecd96e4320ca6e089fea50dd88dd138efc7728b40394e645c167cf30bfe00de9", "e837597389e921bfabccb4624e12765961ff0f28a20786da6f360d2a2be2fa5a", "0653d2d3d0724423c604902ed6aebd51e2efdef2221e90d3a635410da984e7ab", "0526b088281e8c62edb7029cfe9a42f3dcd5ad11943bc20565928378bc4ae0fb", "64ed332f520e76af3430affbbf35fc200b98cb08d6293555684649e2bd84d334", "f7579a22982616d24a0f7349bdac6063ce8b01e3024ef360e84120e360461e58"], "compiled_tokens_by_order": [311, 311, 311, 311, 311, 311], "choice": "ibn_sina", "probs": {"darwin": 0.09346998449346595, "einstein": 0.08248697180011387, "hypatia": 0.13599802496248936, "ibn_sina": 0.5378826228923133, "sagan": 0.05669241135815152, "socrates": 0.09346998449346595}, "confidence": 0.20944849588095504, "seconds": 0.1768614863346253, "seconds_all_orders": 1.061168918007752, "seconds_by_order": [0.1769579520005209, 0.17691418200047337, 0.1770829010019952, 0.1766537050025363, 0.17671612400226877, 0.1768440539999574], "argmax_stable": false, "spread": {"darwin": 0.1452299599937764, "einstein": 0.2604145772984132, "hypatia": 0.1089751967345092, "ibn_sina": 0.3250348150585204, "sagan": 0.03301846597870509, "socrates": 0.27837780213680857}, "spread_2dp": {"darwin": 0.14, "einstein": 0.26, "hypatia": 0.11000000000000001, "ibn_sina": 0.33000000000000007, "sagan": 0.039999999999999994, "socrates": 0.27999999999999997}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.09346998449346595, "einstein": 0.08248697180011387, "hypatia": 0.13599802496248936, "ibn_sina": 0.5378826228923133, "sagan": 0.05669241135815152, "socrates": 0.09346998449346595}, {"ibn_sina": 0.36031173303590286, "hypatia": 0.06261279072111287, "sagan": 0.043033099777803736, "socrates": 0.28061105983817614, "darwin": 0.10323103987979439, "einstein": 0.15020027674720998}, {"hypatia": 0.08669908907489302, "darwin": 0.12614643025918804, "sagan": 0.02483970490336575, "ibn_sina": 0.3429015490985271, "einstein": 0.3429015490985271, "socrates": 0.07651167756549902}, {"ibn_sina": 0.2869364896338642, "einstein": 0.2869364896338642, "darwin": 0.17403577835325484, "sagan": 0.0440031358544014, "hypatia": 0.13553920046395632, "socrates": 0.07254890606065904}, {"einstein": 0.28840769661113, "sagan": 0.023673945379446427, "socrates": 0.19821951780162905, "ibn_sina": 0.32680873526746856, "darwin": 0.10609926225884539, "hypatia": 0.056790842681480574}, {"socrates": 0.35092670819746763, "ibn_sina": 0.21284780783379284, "hypatia": 0.16576603941598977, "einstein": 0.21284780783379284, "sagan": 0.028805818359478465, "darwin": 0.028805818359478465}], "prompt_tokens": 311, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.09346998449346595, "einstein": 0.08248697180011387, "hypatia": 0.13599802496248936, "ibn_sina": 0.5378826228923133, "sagan": 0.05669241135815152, "socrates": 0.09346998449346595}, "confidence": 0.20944849588095504, "choice": "ibn_sina"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 176.93247300121584, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 311, "compiled_prompt_sha256": "ecd96e4320ca6e089fea50dd88dd138efc7728b40394e645c167cf30bfe00de9", "compiled_ids_sha256": "ea289ecc97718490e97f10028ee7f3325d7a73f38345e7ea80efb30da1bcf2a9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 17.5, 18.0, 19.375, 17.125, 17.625], "runtime_probabilities": [0.09346998482942581, 0.08248697966337204, 0.13599802553653717, 0.5378826260566711, 0.05669241398572922, 0.09346998482942581], "decide_seconds": 0.17309347199989134, "probabilities_f64_minus_runtime_max_abs": 7.863258169593657e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.36031173303590286, "hypatia": 0.06261279072111287, "sagan": 0.043033099777803736, "socrates": 0.28061105983817614, "darwin": 0.10323103987979439, "einstein": 0.15020027674720998}, "confidence": 0.13357777562029227, "choice": "ibn_sina"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 176.882943000237, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 311, "compiled_prompt_sha256": "e837597389e921bfabccb4624e12765961ff0f28a20786da6f360d2a2be2fa5a", "compiled_ids_sha256": "5a5e18daf95230b049598d927906f664779fdb31a02062ba877faf9076e9b3d6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.0, 17.25, 16.875, 18.75, 17.75, 18.125], "runtime_probabilities": [0.36031174659729004, 0.06261278688907623, 0.04303310066461563, 0.2806110680103302, 0.1032310500741005, 0.1502002775669098], "decide_seconds": 0.17318402199816774, "probabilities_f64_minus_runtime_max_abs": 1.3561387179894524e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.08669908907489302, "darwin": 0.12614643025918804, "sagan": 0.02483970490336575, "ibn_sina": 0.3429015490985271, "einstein": 0.3429015490985271, "socrates": 0.07651167756549902}, "confidence": 0.1652666850505038, "choice": "ibn_sina"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 177.05975100034266, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 311, "compiled_prompt_sha256": "0653d2d3d0724423c604902ed6aebd51e2efdef2221e90d3a635410da984e7ab", "compiled_ids_sha256": "35556c5e1ad3b0e67bb75b007d39f6a764d42dbd19732cd55ad31e2c38eede96", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 18.0, 16.375, 19.0, 19.0, 17.5], "runtime_probabilities": [0.08669909089803696, 0.12614645063877106, 0.02483970671892166, 0.3429015576839447, 0.3429015576839447, 0.0765116810798645], "decide_seconds": 0.17319237199990312, "probabilities_f64_minus_runtime_max_abs": 2.0379583021146175e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.2869364896338642, "einstein": 0.2869364896338642, "darwin": 0.17403577835325484, "sagan": 0.0440031358544014, "hypatia": 0.13553920046395632, "socrates": 0.07254890606065904}, "confidence": 0.09618041869877259, "choice": "ibn_sina"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 176.63095500029158, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 311, "compiled_prompt_sha256": "0526b088281e8c62edb7029cfe9a42f3dcd5ad11943bc20565928378bc4ae0fb", "compiled_ids_sha256": "ed53035ba24f6c1d397aff1bf0030dd16703f34b3d1bbf11757e2d1f4bc6102f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.75, 18.75, 18.25, 16.875, 18.0, 17.375], "runtime_probabilities": [0.286936491727829, 0.286936491727829, 0.17403578758239746, 0.044003140181303024, 0.1355392038822174, 0.07254891097545624], "decide_seconds": 0.1729160340000817, "probabilities_f64_minus_runtime_max_abs": 9.22914261680674e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.28840769661113, "sagan": 0.023673945379446427, "socrates": 0.19821951780162905, "ibn_sina": 0.32680873526746856, "darwin": 0.10609926225884539, "hypatia": 0.056790842681480574}, "confidence": 0.14361794351669865, "choice": "ibn_sina"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 176.69361499793013, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 311, "compiled_prompt_sha256": "64ed332f520e76af3430affbbf35fc200b98cb08d6293555684649e2bd84d334", "compiled_ids_sha256": "3c389e9fc6ff42665e40d1f441700c2ecc2a8bb6feb63c26baef50b274e2a6c7", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.875, 16.375, 18.5, 19.0, 17.875, 17.25], "runtime_probabilities": [0.28840771317481995, 0.023673947900533676, 0.19821952283382416, 0.32680875062942505, 0.10609926283359528, 0.05679083988070488], "decide_seconds": 0.17299892299706698, "probabilities_f64_minus_runtime_max_abs": 1.6563689919646407e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.35092670819746763, "ibn_sina": 0.21284780783379284, "hypatia": 0.16576603941598977, "einstein": 0.21284780783379284, "sagan": 0.028805818359478465, "darwin": 0.028805818359478465}, "confidence": 0.1469952628598714, "choice": "socrates"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 176.82149399843183, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 311, "compiled_prompt_sha256": "f7579a22982616d24a0f7349bdac6063ce8b01e3024ef360e84120e360461e58", "compiled_ids_sha256": "19c30f7bea095ef8843df96eaa7a7b5c0e2aa3a462a81a3fc19304ab57357dc2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.0, 18.5, 18.25, 18.5, 16.5, 16.5], "runtime_probabilities": [0.3509266972541809, 0.21284779906272888, 0.16576604545116425, 0.21284779906272888, 0.028805818408727646, 0.028805818408727646], "decide_seconds": 0.17307729199819732, "probabilities_f64_minus_runtime_max_abs": 1.094328672124334e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:48Z"} {"id": "c-ibn_sina-01-starter-03", "uid": "c068", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "prompt_sha256_by_order": ["34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf"], "request_sha256": "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "compiled_prompt_sha256_by_order": ["0ad9f34f9bb88bf2b729ba7edfdf37192f81ae6be3b2bb68a023efff32a6297e", "b5975b47ab93a6abc1b31cba510da59f0acc0a5ef0f2ca35a1697e3811c3e734", "ac11a482d4195265088df28225a75a7c7fa4bfb7a2ddf6df2890362c4f4a918f", "8bcad8b709de4d85c01f2b966a8a44a7db39a90c37172f1044258555fad31378", "190de18c143d4fe6df134521b00c7220f889a90dafe55c70e1d39d0cf8de89d7", "2ff33004630d8056e520d0ea2e5bea927888a17c9cf79b258e1aa233284c5b44"], "compiled_tokens_by_order": [371, 371, 371, 371, 371, 371], "choice": "ibn_sina", "probs": {"darwin": 0.0014953281868312453, "einstein": 0.0010277230310830588, "hypatia": 0.0008003915013880039, "ibn_sina": 0.9946050321261992, "sagan": 0.0009069623916456513, "socrates": 0.001164562762852918}, "confidence": 0.9764995648793807, "seconds": 0.1951284376670325, "seconds_all_orders": 1.170770626002195, "seconds_by_order": [0.19540155900176615, 0.195231360001344, 0.19523277999905986, 0.1955397579986311, 0.19464670500019565, 0.1947184640011983], "argmax_stable": true, "spread": {"darwin": 0.001265121366889999, "einstein": 0.002619264962521141, "hypatia": 0.0025100097334447525, "ibn_sina": 0.005985465952313862, "sagan": 0.0007276841850992572, "socrates": 0.0010249352684836065}, "spread_2dp": {"darwin": 0.0, "einstein": 0.0, "hypatia": 0.0, "ibn_sina": 0.010000000000000009, "sagan": 0.0, "socrates": 0.0}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0014953281868312453, "einstein": 0.0010277230310830588, "hypatia": 0.0008003915013880039, "ibn_sina": 0.9946050321261992, "sagan": 0.0009069623916456513, "socrates": 0.001164562762852918}, {"ibn_sina": 0.9984295023567712, "hypatia": 0.0002776715181838432, "sagan": 0.00017927820654639403, "socrates": 0.0004873286910937903, "darwin": 0.00045780293829729977, "einstein": 0.00016841628910745494}, {"hypatia": 0.0027876812516285958, "darwin": 0.0006220157641186638, "sagan": 0.0002938194422473014, "ibn_sina": 0.9924832716693636, "einstein": 0.0027876812516285958, "socrates": 0.0010255306210132347}, {"ibn_sina": 0.9984687376216774, "einstein": 0.00045782092857054045, "darwin": 0.0002302068199412463, "sagan": 0.00021625929380155652, "hypatia": 0.0004873478416399694, "socrates": 0.00013962749436931154}, {"einstein": 0.0008543569864254475, "sagan": 0.00022994706340767813, "socrates": 0.0002605641591876699, "ibn_sina": 0.9973421038485035, "darwin": 0.0009094582779876499, "hypatia": 0.00040356966448799056}, {"socrates": 0.0009666286906246386, "ibn_sina": 0.9958126744077861, "hypatia": 0.000586289937421721, "einstein": 0.0016964861850322077, "sagan": 0.0003138183894306598, "darwin": 0.0006241023897047206}], "prompt_tokens": 371, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0014953281868312453, "einstein": 0.0010277230310830588, "hypatia": 0.0008003915013880039, "ibn_sina": 0.9946050321261992, "sagan": 0.0009069623916456513, "socrates": 0.001164562762852918}, "confidence": 0.9764995648793807, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 195.3743990015937, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "0ad9f34f9bb88bf2b729ba7edfdf37192f81ae6be3b2bb68a023efff32a6297e", "compiled_ids_sha256": "1fa8144179d5b9344965adbb89b17f9bf2ac697e24f6cd4b3cb376295fba968c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.5, 14.125, 13.875, 21.0, 14.0, 14.25], "runtime_probabilities": [0.0014953281497582793, 0.001027722959406674, 0.0008003914845176041, 0.9946050047874451, 0.0009069623774848878, 0.0011645626509562135], "decide_seconds": 0.1908868939972308, "probabilities_f64_minus_runtime_max_abs": 2.733875414939746e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.9984295023567712, "hypatia": 0.0002776715181838432, "sagan": 0.00017927820654639403, "socrates": 0.0004873286910937903, "darwin": 0.00045780293829729977, "einstein": 0.00016841628910745494}, "confidence": 0.9921363328251858, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 195.2056709997123, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "b5975b47ab93a6abc1b31cba510da59f0acc0a5ef0f2ca35a1697e3811c3e734", "compiled_ids_sha256": "be787a016be8e0db152762e222cfa3f08a65b376b275768e4c67f1b759dfa9e9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.625, 12.4375, 12.0, 13.0, 12.9375, 11.9375], "runtime_probabilities": [0.9984295964241028, 0.0002776715555228293, 0.00017927822773344815, 0.0004873287689406425, 0.00045780299114994705, 0.00016841631440911442], "decide_seconds": 0.19080698399920948, "probabilities_f64_minus_runtime_max_abs": 9.406733159078584e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0027876812516285958, "darwin": 0.0006220157641186638, "sagan": 0.0002938194422473014, "ibn_sina": 0.9924832716693636, "einstein": 0.0027876812516285958, "socrates": 0.0010255306210132347}, "confidence": 0.96968033422356, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 195.20789999660337, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "ac11a482d4195265088df28225a75a7c7fa4bfb7a2ddf6df2890362c4f4a918f", "compiled_ids_sha256": "6a98be94c1b57bc9c628b884f2a7055e00de5a20206781ad8e766b970e9d3f0c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.25, 13.75, 13.0, 21.125, 15.25, 14.25], "runtime_probabilities": [0.0027876815292984247, 0.0006220157374627888, 0.0002938194666057825, 0.9924833178520203, 0.0027876815292984247, 0.001025530626066029], "decide_seconds": 0.19068288600101369, "probabilities_f64_minus_runtime_max_abs": 4.6182656698867675e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.9984687376216774, "einstein": 0.00045782092857054045, "darwin": 0.0002302068199412463, "sagan": 0.00021625929380155652, "hypatia": 0.0004873478416399694, "socrates": 0.00013962749436931154}, "confidence": 0.9923204951351221, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 195.51473799947416, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "8bcad8b709de4d85c01f2b966a8a44a7db39a90c37172f1044258555fad31378", "compiled_ids_sha256": "2424748256d78181298665b17bdfc3793da88705ecbf1f6b87a392bf2e91d9f0", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.625, 12.9375, 12.25, 12.1875, 13.0, 11.75], "runtime_probabilities": [0.998468816280365, 0.0004578209773171693, 0.0002302068314747885, 0.0002162593009416014, 0.000487347919261083, 0.00013962751836515963], "decide_seconds": 0.19115606199920876, "probabilities_f64_minus_runtime_max_abs": 7.865868756340433e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.0008543569864254475, "sagan": 0.00022994706340767813, "socrates": 0.0002605641591876699, "ibn_sina": 0.9973421038485035, "darwin": 0.0009094582779876499, "hypatia": 0.00040356966448799056}, "confidence": 0.9875597767740016, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 194.62204499723157, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "190de18c143d4fe6df134521b00c7220f889a90dafe55c70e1d39d0cf8de89d7", "compiled_ids_sha256": "ba54ae93476e974e164bb01ac096bcc1ceace9c5633deb1b1e80d019a3a559cd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.0625, 12.75, 12.875, 21.125, 14.125, 13.3125], "runtime_probabilities": [0.0008543570293113589, 0.0002299470652360469, 0.0002605641493573785, 0.9973421692848206, 0.0009094583219848573, 0.0004035696911159903], "decide_seconds": 0.19025300800058176, "probabilities_f64_minus_runtime_max_abs": 6.543631703159747e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0009666286906246386, "ibn_sina": 0.9958126744077861, "hypatia": 0.000586289937421721, "einstein": 0.0016964861850322077, "sagan": 0.0003138183894306598, "darwin": 0.0006241023897047206}, "confidence": 0.9814647831755213, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 194.694244000857, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "2ff33004630d8056e520d0ea2e5bea927888a17c9cf79b258e1aa233284c5b44", "compiled_ids_sha256": "9e59dfc8370018894c44a9c7a769c1711f03da0724329efe0f2480980e7df9bb", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.3125, 21.25, 13.8125, 14.875, 13.1875, 13.875], "runtime_probabilities": [0.0009666287223808467, 0.995812714099884, 0.0005862899706698954, 0.0016964864917099476, 0.0003138184256386012, 0.0006241023656912148], "decide_seconds": 0.19031397800063132, "probabilities_f64_minus_runtime_max_abs": 3.9692097941745885e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:50Z"} {"id": "c-ibn_sina-01-starter-03", "uid": "c069", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "prompt_sha256_by_order": ["278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442"], "request_sha256": "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "compiled_prompt_sha256_by_order": ["02824b55841c1a917666db92f2dab2f28df3699581fc36e4d28cc000f8f4a105", "3ecad7421cb8e33b4a59fe7c274de9748170c6ee9f704a61867d212b5e75f8e3", "7bbe1ebe15ede610a3ec8498bcdec57585731b432ebb7110a5f114d85cea1a5d", "ab0eb080710e73e679430a3eaba59d08e0a8167fd5d813ba707d610bc197b462", "042751ebc05e9d1454186cdd6852bc9b07fe294fe08da90a89e37231cf9d98cf", "3f9029cec46662495600d1afce7d6d5af2b636cd3b5b845e97cd92ab4ec9e17e"], "compiled_tokens_by_order": [355, 355, 355, 355, 355, 355], "choice": "ibn_sina", "probs": {"darwin": 0.002536243434399052, "einstein": 0.0022382269750576803, "hypatia": 0.019949099878591028, "ibn_sina": 0.9612014606291475, "sagan": 0.01209974071003503, "socrates": 0.001975228372769691}, "confidence": 0.8824282451019971, "seconds": 0.18743488066744854, "seconds_all_orders": 1.1246092840046913, "seconds_by_order": [0.18777914800011786, 0.18731301200023154, 0.18784080800105585, 0.18725527200149372, 0.18725027199980104, 0.18717077200199128], "argmax_stable": true, "spread": {"darwin": 0.0031194117819440723, "einstein": 0.0036566808409063244, "hypatia": 0.1251134299899584, "ibn_sina": 0.11968196319154989, "sagan": 0.010094871076965908, "socrates": 0.002728534079648348}, "spread_2dp": {"darwin": 0.0, "einstein": 0.0, "hypatia": 0.12000000000000001, "ibn_sina": 0.12, "sagan": 0.01, "socrates": 0.0}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.002536243434399052, "einstein": 0.0022382269750576803, "hypatia": 0.019949099878591028, "ibn_sina": 0.9612014606291475, "sagan": 0.01209974071003503, "socrates": 0.001975228372769691}, {"ibn_sina": 0.9814259103300327, "hypatia": 0.007039284452281417, "sagan": 0.0054822002436991325, "socrates": 0.003123663540595446, "darwin": 0.0017798098356694696, "einstein": 0.0011491315977218619}, {"hypatia": 0.015757652338651347, "darwin": 0.0012934668705340715, "sagan": 0.0024165142519477867, "ibn_sina": 0.974891429121295, "einstein": 0.004805812438628186, "socrates": 0.0008351249789436202}, {"ibn_sina": 0.8617439471384828, "einstein": 0.0016635571671226766, "darwin": 0.001468084047258166, "sagan": 0.0025765677439495674, "hypatia": 0.1321527144422398, "socrates": 0.00039512946094709807}, {"einstein": 0.003518655786116249, "sagan": 0.0020048696330691226, "socrates": 0.0009470333572680046, "ibn_sina": 0.9756257282669962, "darwin": 0.003987159360912265, "hypatia": 0.013916553595638181}, {"socrates": 0.0025109108750672273, "ibn_sina": 0.9516007682425374, "hypatia": 0.03689756596066247, "einstein": 0.004690999091816951, "sagan": 0.003432008250947784, "darwin": 0.0008677475789681927}], "prompt_tokens": 355, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.002536243434399052, "einstein": 0.0022382269750576803, "hypatia": 0.019949099878591028, "ibn_sina": 0.9612014606291475, "sagan": 0.01209974071003503, "socrates": 0.001975228372769691}, "confidence": 0.8824282451019971, "choice": "ibn_sina"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.7446479993523, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "02824b55841c1a917666db92f2dab2f28df3699581fc36e4d28cc000f8f4a105", "compiled_ids_sha256": "052585d008f125e9a9c56574e6532af6ea4dffeaa060a8ab415742f7363867e1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.6875, 14.5625, 16.75, 20.625, 16.25, 14.4375], "runtime_probabilities": [0.002536243526265025, 0.0022382268216460943, 0.01994909904897213, 0.9612014889717102, 0.012099741026759148, 0.001975228311493993], "decide_seconds": 0.18348110199804069, "probabilities_f64_minus_runtime_max_abs": 2.8342562741201505e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.9814259103300327, "hypatia": 0.007039284452281417, "sagan": 0.0054822002436991325, "socrates": 0.003123663540595446, "darwin": 0.0017798098356694696, "einstein": 0.0011491315977218619}, "confidence": 0.9336423831999562, "choice": "ibn_sina"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.2875120025128, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "3ecad7421cb8e33b4a59fe7c274de9748170c6ee9f704a61867d212b5e75f8e3", "compiled_ids_sha256": "f95fa995b9c8fb33c4c91228baf13be38103e65eb2ccc52cdc0e9fbefbf430c9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.375, 15.4375, 15.1875, 14.625, 14.0625, 13.625], "runtime_probabilities": [0.9814258813858032, 0.007039283402264118, 0.005482200067490339, 0.0031236635986715555, 0.0017798097105696797, 0.0011491314508020878], "decide_seconds": 0.18316302399762208, "probabilities_f64_minus_runtime_max_abs": 2.8944229457650295e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.015757652338651347, "darwin": 0.0012934668705340715, "sagan": 0.0024165142519477867, "ibn_sina": 0.974891429121295, "einstein": 0.004805812438628186, "socrates": 0.0008351249789436202}, "confidence": 0.9191148479519787, "choice": "ibn_sina"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.81451799804927, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "7bbe1ebe15ede610a3ec8498bcdec57585731b432ebb7110a5f114d85cea1a5d", "compiled_ids_sha256": "dc68a45dcb327e734dfecde766161ceaa2eafbddf62c52cedae022d96a06b1b2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 14.125, 14.75, 20.75, 15.4375, 13.6875], "runtime_probabilities": [0.01575765199959278, 0.0012934667756780982, 0.002416514325886965, 0.9748913645744324, 0.004805812146514654, 0.0008351249271072447], "decide_seconds": 0.1835414609995496, "probabilities_f64_minus_runtime_max_abs": 6.454686263257514e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.8617439471384828, "einstein": 0.0016635571671226766, "darwin": 0.001468084047258166, "sagan": 0.0025765677439495674, "hypatia": 0.1321527144422398, "socrates": 0.00039512946094709807}, "confidence": 0.7575826608395727, "choice": "ibn_sina"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.22977200013702, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "ab0eb080710e73e679430a3eaba59d08e0a8167fd5d813ba707d610bc197b462", "compiled_ids_sha256": "7bfff541c49a042871015082fdcb64cc955c8134ad0d1b79e159a2e7eded1049", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.25, 14.0, 13.875, 14.4375, 18.375, 12.5625], "runtime_probabilities": [0.8617438673973083, 0.001663557137362659, 0.001468084054067731, 0.0025765676982700825, 0.1321527063846588, 0.0003951294347643852], "decide_seconds": 0.18308209400129272, "probabilities_f64_minus_runtime_max_abs": 7.974117444131679e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.003518655786116249, "sagan": 0.0020048696330691226, "socrates": 0.0009470333572680046, "ibn_sina": 0.9756257282669962, "darwin": 0.003987159360912265, "hypatia": 0.013916553595638181}, "confidence": 0.9193426424881571, "choice": "ibn_sina"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.22398299723864, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "042751ebc05e9d1454186cdd6852bc9b07fe294fe08da90a89e37231cf9d98cf", "compiled_ids_sha256": "fe286b5f239c8228fcb9ad0e7d91d3e03e70bebc6af01eb2b5f7934b430bd552", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.125, 14.5625, 13.8125, 20.75, 15.25, 16.5], "runtime_probabilities": [0.0035186561290174723, 0.0020048697479069233, 0.0009470333461649716, 0.97562575340271, 0.003987159579992294, 0.013916553929448128], "decide_seconds": 0.18298959499952616, "probabilities_f64_minus_runtime_max_abs": 2.513571373263801e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0025109108750672273, "ibn_sina": 0.9516007682425374, "hypatia": 0.03689756596066247, "einstein": 0.004690999091816951, "sagan": 0.003432008250947784, "darwin": 0.0008677475789681927}, "confidence": 0.868991607355295, "choice": "ibn_sina"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.14404300044407, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "3f9029cec46662495600d1afce7d6d5af2b636cd3b5b845e97cd92ab4ec9e17e", "compiled_ids_sha256": "a9229864e6acd49900507f8a23c861d2faedba6f3f76c71372044447532b1a1b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.9375, 20.875, 17.625, 15.5625, 15.25, 13.875], "runtime_probabilities": [0.002510911086574197, 0.9516008496284485, 0.03689756989479065, 0.004690999630838633, 0.003432008670642972, 0.0008677476434968412], "decide_seconds": 0.1829607860017859, "probabilities_f64_minus_runtime_max_abs": 8.138591112150095e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:51Z"} {"id": "c-ibn_sina-01-starter-03", "uid": "c070", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "prompt_sha256_by_order": ["1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728"], "request_sha256": "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "compiled_prompt_sha256_by_order": ["646ca9742583b77ea42ed581986481a7143ca23eac7d883e38862c3cb4e2dc56", "dc1c333565974be045f36a600d54c50e1aa495f573dc2b17ace93b9a2c7938bf", "ecd1b862997cce847636020e0e726395ee6c9500788bcaf5b73174dd05960a9a", "02f3ff7f548a223def5e7215730f9572be0e74a03cb59652d7695138f0a8ff85", "be65ec3c9a3389e1d13c475134d8ad45cac3e1c23f304e74cdb267d6d7747916", "834fd24ef1341869a67187414c18d1d9d11e271351083da69d4bd1f327e9c426"], "compiled_tokens_by_order": [371, 371, 371, 371, 371, 371], "choice": "ibn_sina", "probs": {"darwin": 0.012880898490148052, "einstein": 0.010031653830790578, "hypatia": 0.33220275049977716, "ibn_sina": 0.6206364456690738, "sagan": 0.012880898490148052, "socrates": 0.011367353020062247}, "confidence": 0.5137106940722721, "seconds": 0.1949351671670835, "seconds_all_orders": 1.169611003002501, "seconds_by_order": [0.19485931300005177, 0.1952424199989764, 0.19535419900057605, 0.19480942300288007, 0.19452575499963132, 0.1948198930003855], "argmax_stable": false, "spread": {"darwin": 0.02551498257731103, "einstein": 0.08932920343585173, "hypatia": 0.516855044261919, "ibn_sina": 0.5891153270607372, "sagan": 0.020216897392177635, "socrates": 0.027553901471975936}, "spread_2dp": {"darwin": 0.019999999999999997, "einstein": 0.09000000000000001, "hypatia": 0.5199999999999999, "ibn_sina": 0.59, "sagan": 0.02, "socrates": 0.03}, "choices_by_order": ["ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.012880898490148052, "einstein": 0.010031653830790578, "hypatia": 0.33220275049977716, "ibn_sina": 0.6206364456690738, "sagan": 0.012880898490148052, "socrates": 0.011367353020062247}, {"ibn_sina": 0.8592061398287639, "hypatia": 0.0549272108549242, "sagan": 0.02289706805229099, "socrates": 0.015736909388086886, "darwin": 0.02940041734677051, "einstein": 0.017832254529163472}, {"hypatia": 0.1063427790457631, "darwin": 0.014391930122326868, "sagan": 0.007703445076422165, "ibn_sina": 0.7857727600853303, "einstein": 0.07308825191499023, "socrates": 0.012700833755167399}, {"ibn_sina": 0.6773905149126732, "einstein": 0.020455421105459753, "darwin": 0.03372528787769982, "sagan": 0.014058791618937064, "hypatia": 0.24919804408090981, "socrates": 0.005171940404320324}, {"einstein": 0.01201169155407078, "sagan": 0.002680170660113355, "socrates": 0.004703845998480302, "ibn_sina": 0.616082246485055, "darwin": 0.0347569819635014, "hypatia": 0.3297650633387792}, {"socrates": 0.03225774747045624, "ibn_sina": 0.27009081276802666, "hypatia": 0.5717822551168432, "einstein": 0.0993608572666423, "sagan": 0.01726632799184125, "darwin": 0.009241999386190374}], "prompt_tokens": 371, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.012880898490148052, "einstein": 0.010031653830790578, "hypatia": 0.33220275049977716, "ibn_sina": 0.6206364456690738, "sagan": 0.012880898490148052, "socrates": 0.011367353020062247}, "confidence": 0.5137106940722721, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 194.8309129984409, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "646ca9742583b77ea42ed581986481a7143ca23eac7d883e38862c3cb4e2dc56", "compiled_ids_sha256": "823c9e43923cf82515b47c19a4962b220c53c5e9d09cfb9b90c5cd78de0ad4e1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 15.75, 19.25, 19.875, 16.0, 15.875], "runtime_probabilities": [0.012880898080766201, 0.010031653568148613, 0.3322027623653412, 0.6206364631652832, 0.012880898080766201, 0.011367353610694408], "decide_seconds": 0.1903770179997082, "probabilities_f64_minus_runtime_max_abs": 1.7496209370193583e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.8592061398287639, "hypatia": 0.0549272108549242, "sagan": 0.02289706805229099, "socrates": 0.015736909388086886, "darwin": 0.02940041734677051, "einstein": 0.017832254529163472}, "confidence": 0.6556054795264861, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 195.2176100021461, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "dc1c333565974be045f36a600d54c50e1aa495f573dc2b17ace93b9a2c7938bf", "compiled_ids_sha256": "a687eab0099bf1295de853b61258082b62c7a027819d9f9c7c075af9aef4dda9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.875, 17.125, 16.25, 15.875, 16.5, 16.0], "runtime_probabilities": [0.8592061400413513, 0.05492720752954483, 0.022897066548466682, 0.015736909583210945, 0.02940041571855545, 0.017832253128290176], "decide_seconds": 0.19089486399752786, "probabilities_f64_minus_runtime_max_abs": 3.3253793677023857e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.1063427790457631, "darwin": 0.014391930122326868, "sagan": 0.007703445076422165, "ibn_sina": 0.7857727600853303, "einstein": 0.07308825191499023, "socrates": 0.012700833755167399}, "confidence": 0.5686112446835623, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 195.32927900218056, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "ecd1b862997cce847636020e0e726395ee6c9500788bcaf5b73174dd05960a9a", "compiled_ids_sha256": "a2e3a44700654256d50354a44333f6a7dca60e7c3386f60b83a69af8b65084c6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.375, 16.375, 15.75, 20.375, 18.0, 16.25], "runtime_probabilities": [0.10634277760982513, 0.014391929842531681, 0.007703445386141539, 0.7857727408409119, 0.07308824360370636, 0.01270083338022232], "decide_seconds": 0.19090321399926324, "probabilities_f64_minus_runtime_max_abs": 1.9244418392183604e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.6773905149126732, "einstein": 0.020455421105459753, "darwin": 0.03372528787769982, "sagan": 0.014058791618937064, "hypatia": 0.24919804408090981, "socrates": 0.005171940404320324}, "confidence": 0.5026306461471413, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 194.78521299970453, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "02f3ff7f548a223def5e7215730f9572be0e74a03cb59652d7695138f0a8ff85", "compiled_ids_sha256": "d1e1e5f6a028d1c32b396256d6d506bcee4d2a16aac309a950f1455b92efe095", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.75, 16.25, 16.75, 15.875, 18.75, 14.875], "runtime_probabilities": [0.677390456199646, 0.02045542001724243, 0.03372528403997421, 0.014058790169656277, 0.2491980642080307, 0.005171939730644226], "decide_seconds": 0.19045033699876512, "probabilities_f64_minus_runtime_max_abs": 5.871302721782712e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.01201169155407078, "sagan": 0.002680170660113355, "socrates": 0.004703845998480302, "ibn_sina": 0.616082246485055, "darwin": 0.0347569819635014, "hypatia": 0.3297650633387792}, "confidence": 0.5115387847510533, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 194.50117499945918, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "be65ec3c9a3389e1d13c475134d8ad45cac3e1c23f304e74cdb267d6d7747916", "compiled_ids_sha256": "533477d5266939ad3a59208b2c821c0985ab669161926d2be619429ebc2e84a7", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 14.4375, 15.0, 19.875, 17.0, 19.25], "runtime_probabilities": [0.012011691927909851, 0.0026801705826073885, 0.004703845363110304, 0.6160821914672852, 0.03475698083639145, 0.32976505160331726], "decide_seconds": 0.19021390899797552, "probabilities_f64_minus_runtime_max_abs": 5.501776989547125e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.03225774747045624, "ibn_sina": 0.27009081276802666, "hypatia": 0.5717822551168432, "einstein": 0.0993608572666423, "sagan": 0.01726632799184125, "darwin": 0.009241999386190374}, "confidence": 0.37115272361617646, "choice": "hypatia"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 194.7957939992193, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 371, "compiled_prompt_sha256": "834fd24ef1341869a67187414c18d1d9d11e271351083da69d4bd1f327e9c426", "compiled_ids_sha256": "02b9fa41fcedaaecb6d6a8c2bf217b6882610354beef3fc1ecf54b000ca77d04", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 19.125, 19.875, 18.125, 16.375, 15.75], "runtime_probabilities": [0.032257746905088425, 0.27009081840515137, 0.5717822313308716, 0.09936084598302841, 0.017266327515244484, 0.009241999126970768], "decide_seconds": 0.1905173470004229, "probabilities_f64_minus_runtime_max_abs": 2.3785971636769432e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:52Z"} {"id": "c-ibn_sina-01-starter-03", "uid": "c071", "task": "kperm", "label": "ibn_sina", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "prompt_sha256_by_order": ["ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c"], "request_sha256": "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "compiled_prompt_sha256_by_order": ["672de484d211354cd558d5df735a439daa860248a900b0c7bbcb4c54d1e2256c", "a8128f28a15ff483622d517bc12f2ed8a5ab0143b3fa4fe4aa64156368f241cd", "5100a77d8b8aaf9bc8b5eef5d31b99e6b8a3d9209c5a84cd1290982a66bc2a58", "fe997899d73ccbd8dca794dbf8f689c833e0142e0a78d35b1ef88ef4ab9def34", "75e050785f2514cb05d9f062d6aff813c2d8c1e32dbb1a08a292d29024926478", "f29fd6a1c397ef3f4513b5168cd48c0fb88a8c198617030be4be406047fe3896"], "compiled_tokens_by_order": [346, 346, 346, 346, 346, 346], "choice": "hypatia", "probs": {"darwin": 0.007558754284520454, "einstein": 0.007558754284520454, "hypatia": 0.497803987145498, "ibn_sina": 0.43931047675016377, "sagan": 0.03606079987950151, "socrates": 0.011707227655795847}, "confidence": 0.46737886638406323, "seconds": 0.18663955349984462, "seconds_all_orders": 1.1198373209990677, "seconds_by_order": [0.18680797499837354, 0.18673629600016284, 0.18669427699933294, 0.18656462800208828, 0.1865520169994852, 0.18648212799962494], "argmax_stable": false, "spread": {"darwin": 0.013872606725376335, "einstein": 0.06997021332396772, "hypatia": 0.6572072930845521, "ibn_sina": 0.7633646269286581, "sagan": 0.08189359903552729, "socrates": 0.031440115496969366}, "spread_2dp": {"darwin": 0.01, "einstein": 0.07, "hypatia": 0.66, "ibn_sina": 0.77, "sagan": 0.08, "socrates": 0.03}, "choices_by_order": ["hypatia", "ibn_sina", "ibn_sina", "ibn_sina", "ibn_sina", "hypatia"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.007558754284520454, "einstein": 0.007558754284520454, "hypatia": 0.497803987145498, "ibn_sina": 0.43931047675016377, "sagan": 0.03606079987950151, "socrates": 0.011707227655795847}, {"ibn_sina": 0.6664231159371417, "hypatia": 0.14869909658567004, "sagan": 0.09019056115077904, "socrates": 0.03317925323508739, "darwin": 0.018904968630750706, "einstein": 0.04260300446057118}, {"hypatia": 0.306633015140395, "darwin": 0.006774405691565986, "sagan": 0.028521313278409482, "ibn_sina": 0.5728658909512964, "einstein": 0.07752896760848817, "socrates": 0.0076764073298451006}, {"ibn_sina": 0.8463131439177806, "einstein": 0.008832070197536892, "darwin": 0.005032361905374372, "sagan": 0.008296962115251752, "hypatia": 0.1297863241259384, "socrates": 0.0017391377381180265}, {"einstein": 0.031677852310028905, "sagan": 0.031677852310028905, "socrates": 0.0192135886598812, "ibn_sina": 0.6362666722203713, "darwin": 0.01592862446378225, "hypatia": 0.26523541003590745}, {"socrates": 0.02097266950813285, "ibn_sina": 0.08294851698912238, "hypatia": 0.7869936172104904, "einstein": 0.06460036998574024, "sagan": 0.03918210502513124, "darwin": 0.005302721281382841}], "prompt_tokens": 346, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.007558754284520454, "einstein": 0.007558754284520454, "hypatia": 0.497803987145498, "ibn_sina": 0.43931047675016377, "sagan": 0.03606079987950151, "socrates": 0.011707227655795847}, "confidence": 0.46737886638406323, "choice": "hypatia"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 186.7816350022622, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 346, "compiled_prompt_sha256": "672de484d211354cd558d5df735a439daa860248a900b0c7bbcb4c54d1e2256c", "compiled_ids_sha256": "1d298de3fe49f6c758af9b8bfaf818d27668637424e77a80f4a96f9bfce9406e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.4375, 15.4375, 19.625, 19.5, 17.0, 15.875], "runtime_probabilities": [0.007558753713965416, 0.007558753713965416, 0.4978039562702179, 0.4393104612827301, 0.03606079891324043, 0.011707227677106857], "decide_seconds": 0.18259257799945772, "probabilities_f64_minus_runtime_max_abs": 3.087528011036511e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.6664231159371417, "hypatia": 0.14869909658567004, "sagan": 0.09019056115077904, "socrates": 0.03317925323508739, "darwin": 0.018904968630750706, "einstein": 0.04260300446057118}, "confidence": 0.3898147270178729, "choice": "ibn_sina"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 186.7113859989331, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 346, "compiled_prompt_sha256": "a8128f28a15ff483622d517bc12f2ed8a5ab0143b3fa4fe4aa64156368f241cd", "compiled_ids_sha256": "1203a8a4a430230d9427ef458f1b56a6a1d3470965f9dde2d9da5b286cc41170", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.5, 18.0, 17.5, 16.5, 15.9375, 16.75], "runtime_probabilities": [0.6664231419563293, 0.1486991047859192, 0.09019056707620621, 0.033179253339767456, 0.018904970958828926, 0.04260300472378731], "decide_seconds": 0.18249401900175144, "probabilities_f64_minus_runtime_max_abs": 2.601918769329359e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.306633015140395, "darwin": 0.006774405691565986, "sagan": 0.028521313278409482, "ibn_sina": 0.5728658909512964, "einstein": 0.07752896760848817, "socrates": 0.0076764073298451006}, "confidence": 0.4125678717909421, "choice": "ibn_sina"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 186.67053699755343, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 346, "compiled_prompt_sha256": "5100a77d8b8aaf9bc8b5eef5d31b99e6b8a3d9209c5a84cd1290982a66bc2a58", "compiled_ids_sha256": "a1aaf0ac4619ace6345bedd8e90dfa5e04ac6d0a393c5629eb6b287fab5bde90", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.125, 15.3125, 16.75, 19.75, 17.75, 15.4375], "runtime_probabilities": [0.30663302540779114, 0.006774405017495155, 0.028521312400698662, 0.572865903377533, 0.07752896845340729, 0.007676407229155302], "decide_seconds": 0.18255258900171611, "probabilities_f64_minus_runtime_max_abs": 1.2426236595253215e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.8463131439177806, "einstein": 0.008832070197536892, "darwin": 0.005032361905374372, "sagan": 0.008296962115251752, "hypatia": 0.1297863241259384, "socrates": 0.0017391377381180265}, "confidence": 0.7067481725492084, "choice": "ibn_sina"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 186.54147800043575, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 346, "compiled_prompt_sha256": "fe997899d73ccbd8dca794dbf8f689c833e0142e0a78d35b1ef88ef4ab9def34", "compiled_ids_sha256": "846b5ae6d8fd6af7ec5f0eb90889786616d7df0f00cb702678789880ff8121b5", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.25, 15.6875, 15.125, 15.625, 18.375, 14.0625], "runtime_probabilities": [0.8463131785392761, 0.008832070976495743, 0.005032361950725317, 0.008296962827444077, 0.12978632748126984, 0.0017391378059983253], "decide_seconds": 0.18244185900039156, "probabilities_f64_minus_runtime_max_abs": 3.462149555932825e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.031677852310028905, "sagan": 0.031677852310028905, "socrates": 0.0192135886598812, "ibn_sina": 0.6362666722203713, "darwin": 0.01592862446378225, "hypatia": 0.26523541003590745}, "confidence": 0.4417387013055186, "choice": "ibn_sina"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 186.5288379995036, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 346, "compiled_prompt_sha256": "75e050785f2514cb05d9f062d6aff813c2d8c1e32dbb1a08a292d29024926478", "compiled_ids_sha256": "7a38008ababcb3728afc87183afc13ede283248e45f4595fecc5945af733d925", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 16.625, 16.125, 19.625, 15.9375, 18.75], "runtime_probabilities": [0.03167785331606865, 0.03167785331606865, 0.019213588908314705, 0.6362667083740234, 0.01592862606048584, 0.26523539423942566], "decide_seconds": 0.18246735999855446, "probabilities_f64_minus_runtime_max_abs": 3.615365218312405e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.02097266950813285, "ibn_sina": 0.08294851698912238, "hypatia": 0.7869936172104904, "einstein": 0.06460036998574024, "sagan": 0.03918210502513124, "darwin": 0.005302721281382841}, "confidence": 0.5491827494771309, "choice": "hypatia"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 186.45870800173725, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 346, "compiled_prompt_sha256": "f29fd6a1c397ef3f4513b5168cd48c0fb88a8c198617030be4be406047fe3896", "compiled_ids_sha256": "dcb3b20454214f3da3cbe2f735f49a5739f43117058d02cc952bbd3445a4b6d8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 17.75, 20.0, 17.5, 17.0, 15.0], "runtime_probabilities": [0.02097266912460327, 0.08294851332902908, 0.7869936227798462, 0.064600370824337, 0.03918210417032242, 0.005302721168845892], "decide_seconds": 0.1823858099996869, "probabilities_f64_minus_runtime_max_abs": 5.569355754353467e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:53Z"} {"id": "c-sagan-01-dessert-01", "uid": "c072", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "prompt_sha256_by_order": ["66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f"], "request_sha256": "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "compiled_prompt_sha256_by_order": ["63461712757b659f4c1e81c96db955b9576be9a529dd979a61ae859b5e4fe20e", "fef35ee41e526b7e88efc66b7cbdab640adbb8dbcd04a746d68aff571097b885", "99da7fcb270f502a8f50833875aab41f4abac2a90246adecdf432875f9252bfa", "d0dc067a75fd89123503cfa14e73711e76c13003de8aa84a7f7dbd72215d1077", "ca62948d8bf67ae01de708aec872f941b3562a277a11dd49f10b1f4a2ee701dc", "26507bfc7594cde5fde7bf393f257cfff45aab5996b11561a7846383af313423"], "compiled_tokens_by_order": [404, 404, 404, 404, 404, 404], "choice": "sagan", "probs": {"darwin": 0.016780818640530805, "einstein": 0.3370517524056974, "hypatia": 0.008437925516662285, "ibn_sina": 0.005117860530227239, "sagan": 0.6296955738773905, "socrates": 0.0029160690294917756}, "confidence": 0.5475404216776361, "seconds": 0.2129242153338661, "seconds_all_orders": 1.2775452920031967, "seconds_by_order": [0.21323638900139485, 0.2134442280002986, 0.21260963500026264, 0.2126667840020673, 0.21274116299900925, 0.21284709300016402], "argmax_stable": false, "spread": {"darwin": 0.0349172721227387, "einstein": 0.6048967529057113, "hypatia": 0.016586050786620964, "ibn_sina": 0.009539074406850442, "sagan": 0.5856714775327523, "socrates": 0.008756426418343222}, "spread_2dp": {"darwin": 0.03, "einstein": 0.61, "hypatia": 0.02, "ibn_sina": 0.01, "sagan": 0.5800000000000001, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.016780818640530805, "einstein": 0.3370517524056974, "hypatia": 0.008437925516662285, "ibn_sina": 0.005117860530227239, "sagan": 0.6296955738773905, "socrates": 0.0029160690294917756}, {"ibn_sina": 0.007674050581118019, "hypatia": 0.0024913994557927426, "sagan": 0.782774440050941, "socrates": 0.005614455342473002, "darwin": 0.02678506840101123, "einstein": 0.17466058616866395}, {"hypatia": 0.008819201265428378, "darwin": 0.018670249225422396, "sagan": 0.7005959335047305, "ibn_sina": 0.007311374648133559, "einstein": 0.2577348405047052, "socrates": 0.006868400851579945}, {"ibn_sina": 0.011728690535142304, "einstein": 0.3219956456417139, "darwin": 0.04357737190387732, "sagan": 0.6015670632809104, "hypatia": 0.018165751264131787, "socrates": 0.002965477374224336}, {"einstein": 0.43800716026171127, "sagan": 0.4963271360827516, "socrates": 0.011672495447834998, "ibn_sina": 0.011672495447834998, "darwin": 0.0407410122823563, "hypatia": 0.001579700477510822}, {"socrates": 0.00813541085966974, "ibn_sina": 0.002189616128291862, "hypatia": 0.004354571638335732, "einstein": 0.7795573390743753, "sagan": 0.19710296251818873, "darwin": 0.008660099781138618}], "prompt_tokens": 404, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.016780818640530805, "einstein": 0.3370517524056974, "hypatia": 0.008437925516662285, "ibn_sina": 0.005117860530227239, "sagan": 0.6296955738773905, "socrates": 0.0029160690294917756}, "confidence": 0.5475404216776361, "choice": "sagan"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 213.20833999925526, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 404, "compiled_prompt_sha256": "63461712757b659f4c1e81c96db955b9576be9a529dd979a61ae859b5e4fe20e", "compiled_ids_sha256": "82d8ecc5cf442b09faecb1a0ed1d65f9bc9153c91f2b0ea1a85d3471da004e30", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 19.25, 15.5625, 15.0625, 19.875, 14.5], "runtime_probabilities": [0.01678081788122654, 0.33705177903175354, 0.008437925949692726, 0.005117861088365316, 0.6296955943107605, 0.002916069235652685], "decide_seconds": 0.20848273700175923, "probabilities_f64_minus_runtime_max_abs": 2.6626056137679655e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.007674050581118019, "hypatia": 0.0024913994557927426, "sagan": 0.782774440050941, "socrates": 0.005614455342473002, "darwin": 0.02678506840101123, "einstein": 0.17466058616866395}, "confidence": 0.6233642982192732, "choice": "sagan"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 213.41930799826514, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 404, "compiled_prompt_sha256": "fef35ee41e526b7e88efc66b7cbdab640adbb8dbcd04a746d68aff571097b885", "compiled_ids_sha256": "4f9580e184738f4c2a44c36a6721b01b4fa56d35156e11683b66dc296b6ce274", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5, 14.375, 20.125, 15.1875, 16.75, 18.625], "runtime_probabilities": [0.007674050983041525, 0.002491399645805359, 0.7827744483947754, 0.005614455323666334, 0.026785068213939667, 0.1746605932712555], "decide_seconds": 0.20865935599795193, "probabilities_f64_minus_runtime_max_abs": 8.343834401181027e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.008819201265428378, "darwin": 0.018670249225422396, "sagan": 0.7005959335047305, "ibn_sina": 0.007311374648133559, "einstein": 0.2577348405047052, "socrates": 0.006868400851579945}, "confidence": 0.5619127532048758, "choice": "sagan"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 212.58468499945593, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 404, "compiled_prompt_sha256": "99da7fcb270f502a8f50833875aab41f4abac2a90246adecdf432875f9252bfa", "compiled_ids_sha256": "315975be997bff408243c7754221b54a4cff448851180b92ed3c48f639fff0f2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5, 16.25, 19.875, 15.3125, 18.875, 15.25], "runtime_probabilities": [0.008819201029837132, 0.018670247867703438, 0.7005959153175354, 0.007311374414712191, 0.25773486495018005, 0.006868400610983372], "decide_seconds": 0.2079968899997766, "probabilities_f64_minus_runtime_max_abs": 2.4445474866574557e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.011728690535142304, "einstein": 0.3219956456417139, "darwin": 0.04357737190387732, "sagan": 0.6015670632809104, "hypatia": 0.018165751264131787, "socrates": 0.002965477374224336}, "confidence": 0.47014590997479044, "choice": "sagan"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 212.64220499870135, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 404, "compiled_prompt_sha256": "d0dc067a75fd89123503cfa14e73711e76c13003de8aa84a7f7dbd72215d1077", "compiled_ids_sha256": "66a30b322521e26134a99c15a39e62d7cd5ccabbef5ad4b602caf86fa2599a76", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 19.25, 17.25, 19.875, 16.375, 14.5625], "runtime_probabilities": [0.011728690937161446, 0.32199564576148987, 0.043577373027801514, 0.6015670299530029, 0.018165750429034233, 0.0029654770623892546], "decide_seconds": 0.20807733000037842, "probabilities_f64_minus_runtime_max_abs": 3.332790743382219e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.43800716026171127, "sagan": 0.4963271360827516, "socrates": 0.011672495447834998, "ibn_sina": 0.011672495447834998, "darwin": 0.0407410122823563, "hypatia": 0.001579700477510822}, "confidence": 0.4677017441053103, "choice": "sagan"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 212.7160729978641, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 404, "compiled_prompt_sha256": "ca62948d8bf67ae01de708aec872f941b3562a277a11dd49f10b1f4a2ee701dc", "compiled_ids_sha256": "d6e1ca776cdb5d092a1cc96e92ced5e56d48b4b1f8b94519859f603a79983a6e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.625, 19.75, 16.0, 16.0, 17.25, 14.0], "runtime_probabilities": [0.4380071759223938, 0.49632716178894043, 0.011672496795654297, 0.011672496795654297, 0.04074101522564888, 0.0015797006199136376], "decide_seconds": 0.20811282999784453, "probabilities_f64_minus_runtime_max_abs": 2.570618884423226e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.00813541085966974, "ibn_sina": 0.002189616128291862, "hypatia": 0.004354571638335732, "einstein": 0.7795573390743753, "sagan": 0.19710296251818873, "darwin": 0.008660099781138618}, "confidence": 0.6475044698230044, "choice": "einstein"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 212.82282300307998, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 404, "compiled_prompt_sha256": "26507bfc7594cde5fde7bf393f257cfff45aab5996b11561a7846383af313423", "compiled_ids_sha256": "525bba058f24dbe6ae5fb76c69670ab1fda2cbeb296202081b3971070bf23601", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5625, 14.25, 14.9375, 20.125, 18.75, 15.625], "runtime_probabilities": [0.008135410957038403, 0.0021896162070333958, 0.004354571457952261, 0.7795572876930237, 0.19710294902324677, 0.008660100400447845], "decide_seconds": 0.20827122800255893, "probabilities_f64_minus_runtime_max_abs": 5.138135161786295e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:54Z"} {"id": "c-sagan-01-dessert-01", "uid": "c073", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "prompt_sha256_by_order": ["985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549"], "request_sha256": "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "compiled_prompt_sha256_by_order": ["929aa8fbc3eee78c7018d5c8c61cbd18f10472b9ec10356a479f5560f36dc969", "c26c7a1876342961a9b16ea85b6008c8b4e2186347b4446ab8e8521fd9f47803", "937544d454a03b10ebdf949ac5cb53a99a436501a7429042faf1e639a219d52b", "352878a3b71191b84b3b4ab58ab21aea718d624cc86267f5f3136bf77b25d176", "aa75f9e6afad585ff0e378f8cf32e706fb131e8b9a464c07e76d9ff215412a7a", "855f48f09ae5a5443c7bc7e73391c1f4b3f2c7afa08c7a3c7b95df727bc83f3d"], "compiled_tokens_by_order": [309, 309, 309, 309, 309, 309], "choice": "sagan", "probs": {"darwin": 0.02711509436690372, "einstein": 0.023928986792081542, "hypatia": 0.03072542723661796, "ibn_sina": 0.012032250187908166, "sagan": 0.8979286048626086, "socrates": 0.008269636553879944}, "confidence": 0.7300634089758582, "seconds": 0.17678651350009508, "seconds_all_orders": 1.0607190810005704, "seconds_by_order": [0.17738903900317382, 0.1767257140018046, 0.1767019539984176, 0.17664002499805065, 0.176814252998156, 0.17644809600096778], "argmax_stable": true, "spread": {"darwin": 0.04003689769532227, "einstein": 0.025151258040907655, "hypatia": 0.028817091922015772, "ibn_sina": 0.010697124993501589, "sagan": 0.06393282191263205, "socrates": 0.022491598895138725}, "spread_2dp": {"darwin": 0.04, "einstein": 0.03, "hypatia": 0.03, "ibn_sina": 0.01, "sagan": 0.06999999999999995, "socrates": 0.03}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.02711509436690372, "einstein": 0.023928986792081542, "hypatia": 0.03072542723661796, "ibn_sina": 0.012032250187908166, "sagan": 0.8979286048626086, "socrates": 0.008269636553879944}, {"ibn_sina": 0.006257190603480111, "hypatia": 0.001908335314602188, "sagan": 0.928649424569441, "socrates": 0.006660744725797475, "darwin": 0.03177663587997771, "einstein": 0.024747668906701544}, {"hypatia": 0.005680086321237391, "darwin": 0.005335947288078545, "sagan": 0.9552436415824866, "ibn_sina": 0.006851491939998879, "einstein": 0.022465177193305576, "socrates": 0.004423655674892855}, {"ibn_sina": 0.01077699530514861, "einstein": 0.014730406074928993, "darwin": 0.045372844983400815, "sagan": 0.9113379532241673, "hypatia": 0.012999537734938145, "socrates": 0.004782262677416122}, {"einstein": 0.020961621335706802, "sagan": 0.8913108196698546, "socrates": 0.02691525457003158, "ibn_sina": 0.014406697610107383, "darwin": 0.03916146432165926, "hypatia": 0.00724414249264048}, {"socrates": 0.01073400427802115, "ibn_sina": 0.003709572616605793, "hypatia": 0.016625151051467557, "einstein": 0.03988166411583665, "sagan": 0.9077024914317116, "darwin": 0.021347116506357275}], "prompt_tokens": 309, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.02711509436690372, "einstein": 0.023928986792081542, "hypatia": 0.03072542723661796, "ibn_sina": 0.012032250187908166, "sagan": 0.8979286048626086, "socrates": 0.008269636553879944}, "confidence": 0.7300634089758582, "choice": "sagan"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 177.3634490018594, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 309, "compiled_prompt_sha256": "929aa8fbc3eee78c7018d5c8c61cbd18f10472b9ec10356a479f5560f36dc969", "compiled_ids_sha256": "9bfb6bfa3765f21c286543e6595240c7937c405557de6095941783fa77c32cf4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 16.5, 16.75, 15.8125, 20.125, 15.4375], "runtime_probabilities": [0.027115095406770706, 0.023928986862301826, 0.030725426971912384, 0.012032250873744488, 0.8979285955429077, 0.00826963596045971], "decide_seconds": 0.17340187000081642, "probabilities_f64_minus_runtime_max_abs": 9.319700899901306e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.006257190603480111, "hypatia": 0.001908335314602188, "sagan": 0.928649424569441, "socrates": 0.006660744725797475, "darwin": 0.03177663587997771, "einstein": 0.024747668906701544}, "confidence": 0.8063570268824524, "choice": "sagan"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 176.7024440014211, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 309, "compiled_prompt_sha256": "c26c7a1876342961a9b16ea85b6008c8b4e2186347b4446ab8e8521fd9f47803", "compiled_ids_sha256": "ba40f96a6ae1e1245d8ff0ffb55fbf494afc96c33709a38394fb5b4f98ec8236", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.25, 14.0625, 20.25, 15.3125, 16.875, 16.625], "runtime_probabilities": [0.006257191300392151, 0.001908335485495627, 0.9286494851112366, 0.00666074501350522, 0.031776636838912964, 0.02474767155945301], "decide_seconds": 0.17299028300112695, "probabilities_f64_minus_runtime_max_abs": 6.054179557679618e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.005680086321237391, "darwin": 0.005335947288078545, "sagan": 0.9552436415824866, "ibn_sina": 0.006851491939998879, "einstein": 0.022465177193305576, "socrates": 0.004423655674892855}, "confidence": 0.8635808392276259, "choice": "sagan"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 176.67911399985314, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 309, "compiled_prompt_sha256": "937544d454a03b10ebdf949ac5cb53a99a436501a7429042faf1e639a219d52b", "compiled_ids_sha256": "105d7187ab5cfda37fbe5608beea109a687bf1cc2508d8b7b48f04a8116bcba6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.375, 15.3125, 20.5, 15.5625, 16.75, 15.125], "runtime_probabilities": [0.005680085625499487, 0.005335947033017874, 0.9552435874938965, 0.006851491052657366, 0.02246517688035965, 0.004423655569553375], "decide_seconds": 0.17294149399822345, "probabilities_f64_minus_runtime_max_abs": 5.40885901623156e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.01077699530514861, "einstein": 0.014730406074928993, "darwin": 0.045372844983400815, "sagan": 0.9113379532241673, "hypatia": 0.012999537734938145, "socrates": 0.004782262677416122}, "confidence": 0.766765138241691, "choice": "sagan"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 176.617545002955, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 309, "compiled_prompt_sha256": "352878a3b71191b84b3b4ab58ab21aea718d624cc86267f5f3136bf77b25d176", "compiled_ids_sha256": "cee15b0c067a810ecb8495956276aa604b36bf64a34a6106272a28249f816cd4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.8125, 16.125, 17.25, 20.25, 16.0, 15.0], "runtime_probabilities": [0.010776994749903679, 0.014730405993759632, 0.045372843742370605, 0.9113379716873169, 0.012999537400901318, 0.004782262723892927], "decide_seconds": 0.17289794400130631, "probabilities_f64_minus_runtime_max_abs": 1.8463149564773573e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.020961621335706802, "sagan": 0.8913108196698546, "socrates": 0.02691525457003158, "ibn_sina": 0.014406697610107383, "darwin": 0.03916146432165926, "hypatia": 0.00724414249264048}, "confidence": 0.7184102064032989, "choice": "sagan"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 176.79255300026853, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 309, "compiled_prompt_sha256": "aa75f9e6afad585ff0e378f8cf32e706fb131e8b9a464c07e76d9ff215412a7a", "compiled_ids_sha256": "b311a78c0a73ff1143291dedba188f07cbc6db0a8fb5cf901e2cdd35b8695325", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 20.375, 16.875, 16.25, 17.25, 15.5625], "runtime_probabilities": [0.020961621776223183, 0.8913108110427856, 0.026915254071354866, 0.014406696893274784, 0.03916146233677864, 0.0072441427037119865], "decide_seconds": 0.17313168200053042, "probabilities_f64_minus_runtime_max_abs": 8.627068948108274e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.01073400427802115, "ibn_sina": 0.003709572616605793, "hypatia": 0.016625151051467557, "einstein": 0.03988166411583665, "sagan": 0.9077024914317116, "darwin": 0.021347116506357275}, "confidence": 0.7566325731720277, "choice": "sagan"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 176.42587600130355, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 309, "compiled_prompt_sha256": "855f48f09ae5a5443c7bc7e73391c1f4b3f2c7afa08c7a3c7b95df727bc83f3d", "compiled_ids_sha256": "b5b74e3ad737b5e5abade785558e11637266442154e1541c89bb8c4d990b96d1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.6875, 14.625, 16.125, 17.0, 20.125, 16.375], "runtime_probabilities": [0.010734003968536854, 0.003709572833031416, 0.01662515103816986, 0.039881665259599686, 0.9077025055885315, 0.02134711667895317], "decide_seconds": 0.17275471500033746, "probabilities_f64_minus_runtime_max_abs": 1.4156819938548892e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:55Z"} {"id": "c-sagan-01-dessert-01", "uid": "c074", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "prompt_sha256_by_order": ["5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d"], "request_sha256": "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "compiled_prompt_sha256_by_order": ["c9162970d7a87671df926d763cd3c86cd736e6a98d9885f5541c96e5393b7873", "3cc5db792d343ae5d2f8338f11575cf3f6c37baa6dd8f862c962baf5496a6a64", "0d8b8e4fe95640102c23b96ccfdf039f3672b8aa493fe9554fe94967a2ca0c2b", "1c62b88c1ff6492b1059799160ba5c6d7e5e163c8c816ceab82eb21115459a88", "744cb9c68e7eec0166f194627adc5e0c0e5c6f6cd7ffd276ee8f6443bce4d387", "1aa36fbb6f910ed5794a5ea13797323b27841063359461e1f7e5ac5680285799"], "compiled_tokens_by_order": [288, 288, 288, 288, 288, 288], "choice": "sagan", "probs": {"darwin": 0.08900992969111375, "einstein": 0.01986079986677869, "hypatia": 0.14675256438526213, "ibn_sina": 0.061175570382640716, "sagan": 0.6576993638495124, "socrates": 0.02550177182469235}, "confidence": 0.3777950619825561, "seconds": 0.17397627216814726, "seconds_all_orders": 1.0438576330088836, "seconds_by_order": [0.17436483200071962, 0.17363060800198582, 0.17357292900123866, 0.17363499800194404, 0.17427185300039127, 0.17438241300260415], "argmax_stable": true, "spread": {"darwin": 0.16506589457030557, "einstein": 0.05282257563462357, "hypatia": 0.13882050573723603, "ibn_sina": 0.07436809584138952, "sagan": 0.44300998055044877, "socrates": 0.11456592301978881}, "spread_2dp": {"darwin": 0.17, "einstein": 0.05, "hypatia": 0.13999999999999999, "ibn_sina": 0.06999999999999999, "sagan": 0.44, "socrates": 0.12000000000000001}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.08900992969111375, "einstein": 0.01986079986677869, "hypatia": 0.14675256438526213, "ibn_sina": 0.061175570382640716, "sagan": 0.6576993638495124, "socrates": 0.02550177182469235}, {"ibn_sina": 0.02600821902676391, "hypatia": 0.007932058648026109, "sagan": 0.8612739277119476, "socrates": 0.02947117315720072, "darwin": 0.055059400111725276, "einstein": 0.02025522134433634}, {"hypatia": 0.03250800428622152, "darwin": 0.03250800428622152, "sagan": 0.8383924805712221, "ibn_sina": 0.03250800428622152, "einstein": 0.041741103749302476, "socrates": 0.02234240282081102}, {"ibn_sina": 0.05091492731720417, "einstein": 0.021224499433234192, "darwin": 0.13840112172366917, "sagan": 0.7028588913285592, "hypatia": 0.06537606076409916, "socrates": 0.021224499433234192}, {"einstein": 0.07268337550140226, "sagan": 0.4182639471614988, "socrates": 0.135790422453023, "ibn_sina": 0.09332730151445962, "darwin": 0.1975738988565271, "hypatia": 0.08236105451308924}, {"socrates": 0.10910269570119216, "ibn_sina": 0.01895920567307009, "hypatia": 0.12362955085922658, "einstein": 0.04548077008236867, "sagan": 0.6278426646415193, "darwin": 0.07498511304262326}], "prompt_tokens": 288, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.08900992969111375, "einstein": 0.01986079986677869, "hypatia": 0.14675256438526213, "ibn_sina": 0.061175570382640716, "sagan": 0.6576993638495124, "socrates": 0.02550177182469235}, "confidence": 0.3777950619825561, "choice": "sagan"}}, "usage": {"input_tokens": 288, "output_tokens": 1}, "latency_ms": 174.33954200168955, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 288, "compiled_prompt_sha256": "c9162970d7a87671df926d763cd3c86cd736e6a98d9885f5541c96e5393b7873", "compiled_ids_sha256": "2e18cf1277f24da4cbb35e3c4d9ff0025b081d2a66b6e030656f384eefb90cf3", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 16.125, 18.125, 17.25, 19.625, 16.375], "runtime_probabilities": [0.08900992572307587, 0.019860800355672836, 0.14675256609916687, 0.06117556616663933, 0.6576993465423584, 0.02550177089869976], "decide_seconds": 0.17053949299952365, "probabilities_f64_minus_runtime_max_abs": 1.7307154043244566e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.02600821902676391, "hypatia": 0.007932058648026109, "sagan": 0.8612739277119476, "socrates": 0.02947117315720072, "darwin": 0.055059400111725276, "einstein": 0.02025522134433634}, "confidence": 0.6626842965581377, "choice": "sagan"}}, "usage": {"input_tokens": 288, "output_tokens": 1}, "latency_ms": 173.60832800113712, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 288, "compiled_prompt_sha256": "3cc5db792d343ae5d2f8338f11575cf3f6c37baa6dd8f862c962baf5496a6a64", "compiled_ids_sha256": "5fe74006f150d870c016d10046b38516d94b44f599a303416c2c59649b0f4470", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 15.4375, 20.125, 16.75, 17.375, 16.375], "runtime_probabilities": [0.02600821852684021, 0.007932058535516262, 0.8612738847732544, 0.029471172019839287, 0.05505939573049545, 0.02025522105395794], "decide_seconds": 0.1700714260005043, "probabilities_f64_minus_runtime_max_abs": 4.2938693156635566e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.03250800428622152, "darwin": 0.03250800428622152, "sagan": 0.8383924805712221, "ibn_sina": 0.03250800428622152, "einstein": 0.041741103749302476, "socrates": 0.02234240282081102}, "confidence": 0.6096370581452822, "choice": "sagan"}}, "usage": {"input_tokens": 288, "output_tokens": 1}, "latency_ms": 173.55090899945935, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 288, "compiled_prompt_sha256": "0d8b8e4fe95640102c23b96ccfdf039f3672b8aa493fe9554fe94967a2ca0c2b", "compiled_ids_sha256": "b36de7ee95b3245e3e669275c594732d3d1489a2a98f64d7a1e1f83d650b3049", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 16.875, 20.125, 16.875, 17.125, 16.5], "runtime_probabilities": [0.0325080044567585, 0.0325080044567585, 0.8383924961090088, 0.0325080044567585, 0.04174110293388367, 0.02234240248799324], "decide_seconds": 0.17005548600354814, "probabilities_f64_minus_runtime_max_abs": 1.5537786723918146e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.05091492731720417, "einstein": 0.021224499433234192, "darwin": 0.13840112172366917, "sagan": 0.7028588913285592, "hypatia": 0.06537606076409916, "socrates": 0.021224499433234192}, "confidence": 0.4335220236425513, "choice": "sagan"}}, "usage": {"input_tokens": 288, "output_tokens": 1}, "latency_ms": 173.61403800168773, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 288, "compiled_prompt_sha256": "1c62b88c1ff6492b1059799160ba5c6d7e5e163c8c816ceab82eb21115459a88", "compiled_ids_sha256": "2fc342ca47a1573e492cd12a6123939e4eca316384c14f971ac9a1ccc3ed674d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 16.125, 18.0, 19.625, 17.25, 16.125], "runtime_probabilities": [0.05091492831707001, 0.021224498748779297, 0.1384011059999466, 0.7028588652610779, 0.06537605822086334, 0.021224498748779297], "decide_seconds": 0.17009746600160724, "probabilities_f64_minus_runtime_max_abs": 2.6067481284641758e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.07268337550140226, "sagan": 0.4182639471614988, "socrates": 0.135790422453023, "ibn_sina": 0.09332730151445962, "darwin": 0.1975738988565271, "hypatia": 0.08236105451308924}, "confidence": 0.12175108176595983, "choice": "sagan"}}, "usage": {"input_tokens": 288, "output_tokens": 1}, "latency_ms": 174.25024299882352, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 288, "compiled_prompt_sha256": "744cb9c68e7eec0166f194627adc5e0c0e5c6f6cd7ffd276ee8f6443bce4d387", "compiled_ids_sha256": "9bcdfb63ad98fa9d39ea67de012cd2af2e9c05cec026eef1080ae7ac37f105c7", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 19.125, 18.0, 17.625, 18.375, 17.5], "runtime_probabilities": [0.07268337160348892, 0.4182639420032501, 0.13579042255878448, 0.09332730621099472, 0.19757390022277832, 0.08236104995012283], "decide_seconds": 0.1707250910003495, "probabilities_f64_minus_runtime_max_abs": 5.1582486615409096e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.10910269570119216, "ibn_sina": 0.01895920567307009, "hypatia": 0.12362955085922658, "einstein": 0.04548077008236867, "sagan": 0.6278426646415193, "darwin": 0.07498511304262326}, "confidence": 0.3289383744169435, "choice": "sagan"}}, "usage": {"input_tokens": 288, "output_tokens": 1}, "latency_ms": 174.36014299892122, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 288, "compiled_prompt_sha256": "1aa36fbb6f910ed5794a5ea13797323b27841063359461e1f7e5ac5680285799", "compiled_ids_sha256": "6b1b395d8c3d0dd87a4da79997213d77240cb646d1f13e095909fe1ed8d00e6a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.875, 16.125, 18.0, 17.0, 19.625, 17.5], "runtime_probabilities": [0.10910268872976303, 0.018959207460284233, 0.12362954765558243, 0.045480772852897644, 0.6278426647186279, 0.07498511672019958], "decide_seconds": 0.17079708000164828, "probabilities_f64_minus_runtime_max_abs": 6.971429128421747e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:56Z"} {"id": "c-sagan-01-dessert-03", "uid": "c075", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "prompt_sha256_by_order": ["d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d"], "request_sha256": "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "compiled_prompt_sha256_by_order": ["efc2df06e0745877c07b202c66d4cf238051ae0a1ae157bb1d319d831d699748", "f20d1844845c6c8e4f2a196816988ad19fff335abd61838ca0450870e8b6ae25", "64b37548a0164dae355cd9c73b2d5501f8160f509954d4eb2eb51fe8ec653271", "ad23a3a76e6a573aa4f2b8864ce42d93dc025e5a1d43171c72356e91eddc1a17", "26ba57f0d7982bb838d8721008767af0625af79cf648c0800a20556c8564704e", "f83db599e2edc116912d3740c5169e74befbc33c37d53e395ea2b3a2388f1c99"], "compiled_tokens_by_order": [327, 327, 327, 327, 327, 327], "choice": "einstein", "probs": {"darwin": 0.0026573452175330877, "einstein": 0.9460804293573352, "hypatia": 0.001107746294363868, "ibn_sina": 0.0013361970889456958, "sagan": 0.047102571017911724, "socrates": 0.0017157110239104441}, "confidence": 0.8663750240351824, "seconds": 0.18467405383307778, "seconds_all_orders": 1.1080443229984667, "seconds_by_order": [0.18526459799977602, 0.18437925499893026, 0.18455200299649732, 0.1847470210013853, 0.1845446229999652, 0.18455682300191256], "argmax_stable": false, "spread": {"darwin": 0.0057142414264512095, "einstein": 0.8424050979277641, "hypatia": 0.004395986617417804, "ibn_sina": 0.004158394181145981, "sagan": 0.8209603510895055, "socrates": 0.012796576398548833}, "spread_2dp": {"darwin": 0.01, "einstein": 0.85, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.82, "socrates": 0.01}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "sagan", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0026573452175330877, "einstein": 0.9460804293573352, "hypatia": 0.001107746294363868, "ibn_sina": 0.0013361970889456958, "sagan": 0.047102571017911724, "socrates": 0.0017157110239104441}, {"ibn_sina": 0.0049930905080679535, "hypatia": 0.0026726087580730635, "sagan": 0.2405805734903875, "socrates": 0.00469057444708901, "darwin": 0.006022816808928339, "einstein": 0.7410403359874542}, {"hypatia": 0.005503732911781672, "darwin": 0.0027674507047617796, "sagan": 0.059170705107207164, "ibn_sina": 0.0029459359405472393, "einstein": 0.9255855583198948, "socrates": 0.00402661701580748}, {"ibn_sina": 0.003355321898911485, "einstein": 0.9303377682216875, "darwin": 0.0017959743927524123, "sagan": 0.05947450372223411, "hypatia": 0.0038020778192927947, "socrates": 0.0012343539451218598}, {"einstein": 0.10367533142957105, "sagan": 0.8680629221074172, "socrates": 0.014030930343670693, "ibn_sina": 0.0054945912700916765, "darwin": 0.007510215819203621, "hypatia": 0.0012260090300457202}, {"socrates": 0.0055187933656912404, "ibn_sina": 0.0013953694839112898, "hypatia": 0.00190724395256605, "einstein": 0.7228192878168938, "sagan": 0.2659103556700188, "darwin": 0.002448949710918797}], "prompt_tokens": 327, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0026573452175330877, "einstein": 0.9460804293573352, "hypatia": 0.001107746294363868, "ibn_sina": 0.0013361970889456958, "sagan": 0.047102571017911724, "socrates": 0.0017157110239104441}, "confidence": 0.8663750240351824, "choice": "einstein"}}, "usage": {"input_tokens": 327, "output_tokens": 1}, "latency_ms": 185.23467799968785, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 327, "compiled_prompt_sha256": "efc2df06e0745877c07b202c66d4cf238051ae0a1ae157bb1d319d831d699748", "compiled_ids_sha256": "6a8b4cc8d5bca66563e796c67461a643efc58f4871ab7302ea9ea36cd395825a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.75, 20.625, 13.875, 14.0625, 17.625, 14.3125], "runtime_probabilities": [0.002657345263287425, 0.9460803866386414, 0.001107746153138578, 0.0013361970195546746, 0.04710256680846214, 0.0017157108522951603], "decide_seconds": 0.18098777100021834, "probabilities_f64_minus_runtime_max_abs": 4.27186938045665e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0049930905080679535, "hypatia": 0.0026726087580730635, "sagan": 0.2405805734903875, "socrates": 0.00469057444708901, "darwin": 0.006022816808928339, "einstein": 0.7410403359874542}, "confidence": 0.6299262858452774, "choice": "einstein"}}, "usage": {"input_tokens": 327, "output_tokens": 1}, "latency_ms": 184.35413499901188, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 327, "compiled_prompt_sha256": "f20d1844845c6c8e4f2a196816988ad19fff335abd61838ca0450870e8b6ae25", "compiled_ids_sha256": "8a9d223ee987f87922beb172dc0b7755174f1097c7ab839b8802b411db25d650", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0, 14.375, 18.875, 14.9375, 15.1875, 20.0], "runtime_probabilities": [0.004993090406060219, 0.002672608708962798, 0.24058055877685547, 0.004690574016422033, 0.006022816989570856, 0.7410402894020081], "decide_seconds": 0.1803579159968649, "probabilities_f64_minus_runtime_max_abs": 4.6585446167313194e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.005503732911781672, "darwin": 0.0027674507047617796, "sagan": 0.059170705107207164, "ibn_sina": 0.0029459359405472393, "einstein": 0.9255855583198948, "socrates": 0.00402661701580748}, "confidence": 0.8196329080984579, "choice": "einstein"}}, "usage": {"input_tokens": 327, "output_tokens": 1}, "latency_ms": 184.52742299996316, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 327, "compiled_prompt_sha256": "64b37548a0164dae355cd9c73b2d5501f8160f509954d4eb2eb51fe8ec653271", "compiled_ids_sha256": "e23eaa8b9eaf5627189988d8e7cc1eb79d9b86d8acf644502e198e33c739a3e1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0, 14.3125, 17.375, 14.375, 20.125, 14.6875], "runtime_probabilities": [0.0055037327110767365, 0.00276745087467134, 0.05917070433497429, 0.0029459360521286726, 0.9255855679512024, 0.004026616923511028], "decide_seconds": 0.18053942400001688, "probabilities_f64_minus_runtime_max_abs": 9.631307640312059e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.003355321898911485, "einstein": 0.9303377682216875, "darwin": 0.0017959743927524123, "sagan": 0.05947450372223411, "hypatia": 0.0038020778192927947, "socrates": 0.0012343539451218598}, "confidence": 0.8353852701955752, "choice": "einstein"}}, "usage": {"input_tokens": 327, "output_tokens": 1}, "latency_ms": 184.72201099939412, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 327, "compiled_prompt_sha256": "ad23a3a76e6a573aa4f2b8864ce42d93dc025e5a1d43171c72356e91eddc1a17", "compiled_ids_sha256": "f479e57cbfa5839bfebcc69460d11a8ac7bede7572e2a3784529cb17fdacb7fa", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.875, 20.5, 14.25, 17.75, 15.0, 13.875], "runtime_probabilities": [0.003355321940034628, 0.9303377270698547, 0.0017959743272513151, 0.05947450175881386, 0.0038020778447389603, 0.0012343538692221045], "decide_seconds": 0.1807991530004074, "probabilities_f64_minus_runtime_max_abs": 4.115183271924394e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.10367533142957105, "sagan": 0.8680629221074172, "socrates": 0.014030930343670693, "ibn_sina": 0.0054945912700916765, "darwin": 0.007510215819203621, "hypatia": 0.0012260090300457202}, "confidence": 0.725848067701406, "choice": "sagan"}}, "usage": {"input_tokens": 327, "output_tokens": 1}, "latency_ms": 184.5202229997085, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 327, "compiled_prompt_sha256": "26ba57f0d7982bb838d8721008767af0625af79cf648c0800a20556c8564704e", "compiled_ids_sha256": "944d31185c967912debb5165ee772499c86190aeb63ef728cd1e04b9014126aa", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.0, 20.125, 16.0, 15.0625, 15.375, 13.5625], "runtime_probabilities": [0.10367532819509506, 0.8680628538131714, 0.014030929654836655, 0.005494590848684311, 0.007510215509682894, 0.0012260089861229062], "decide_seconds": 0.18063748399799806, "probabilities_f64_minus_runtime_max_abs": 6.829424581411558e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0055187933656912404, "ibn_sina": 0.0013953694839112898, "hypatia": 0.00190724395256605, "einstein": 0.7228192878168938, "sagan": 0.2659103556700188, "darwin": 0.002448949710918797}, "confidence": 0.6364555388162472, "choice": "einstein"}}, "usage": {"input_tokens": 327, "output_tokens": 1}, "latency_ms": 184.53159300042898, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 327, "compiled_prompt_sha256": "f83db599e2edc116912d3740c5169e74befbc33c37d53e395ea2b3a2388f1c99", "compiled_ids_sha256": "043c2fcc4ed7ad12eb2a90c0d51efe2d25c83ad18257b651a77263792833aeb8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.875, 13.5, 13.8125, 19.75, 18.75, 14.0625], "runtime_probabilities": [0.005518792662769556, 0.0013953694142401218, 0.0019072438590228558, 0.7228192687034607, 0.2659103572368622, 0.0024489497300237417], "decide_seconds": 0.18056632400111994, "probabilities_f64_minus_runtime_max_abs": 1.9113433058493e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:58Z"} {"id": "c-sagan-01-dessert-03", "uid": "c076", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "prompt_sha256_by_order": ["16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac"], "request_sha256": "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "compiled_prompt_sha256_by_order": ["d1ab3fa14e7acc3c64152e9b9fefb12997ba7cb18617a7d4ba7df9edddd170aa", "a67d01f47b4e062017ef0625a80f0841b27f2ee07b838a1444e1364e594abbec", "ffcbf6fb24313235b7d591c15b4afb758fa9a761ccf6d0770a2a4294ee2bcaf7", "e061af839a84e5f40968432bbf43b1b0645d1b6f93e35f31dfdee2362b65e1b5", "067e4724717953ff29f69852ea64f6abec0a34d07cfeaa8a1edce8084c915f87", "27dec1ce00447b2c51aa4dbbfa98f2d68b019252e4fcb2ce695c4187f86b8845"], "compiled_tokens_by_order": [409, 409, 409, 409, 409, 409], "choice": "einstein", "probs": {"darwin": 0.014556269484544377, "einstein": 0.48203744281320826, "hypatia": 0.011336434073161554, "ibn_sina": 0.008293912344448233, "sagan": 0.48203744281320826, "socrates": 0.0017384984714294077}, "confidence": 0.5163049815451398, "seconds": 0.2141616508330723, "seconds_all_orders": 1.2849699049984338, "seconds_by_order": [0.21487875700040604, 0.21397742400222342, 0.21424204199865926, 0.21419258199966862, 0.2135747469983471, 0.21410435299912933], "argmax_stable": false, "spread": {"darwin": 0.012117681508238295, "einstein": 0.6625525301884093, "hypatia": 0.010792296562912144, "ibn_sina": 0.010847980554651174, "sagan": 0.6481606032800116, "socrates": 0.005929725485635446}, "spread_2dp": {"darwin": 0.01, "einstein": 0.67, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.65, "socrates": 0.01}, "choices_by_order": ["einstein", "sagan", "einstein", "sagan", "sagan", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.014556269484544377, "einstein": 0.48203744281320826, "hypatia": 0.011336434073161554, "ibn_sina": 0.008293912344448233, "sagan": 0.48203744281320826, "socrates": 0.0017384984714294077}, {"ibn_sina": 0.010178451945149157, "hypatia": 0.002573515675729548, "sagan": 0.7135645706656287, "socrates": 0.002739493176752355, "darwin": 0.008438233040528586, "einstein": 0.26250573549621165}, {"hypatia": 0.004555882128576662, "darwin": 0.002438587976306082, "sagan": 0.21951469405101837, "ibn_sina": 0.006227152548854217, "einstein": 0.7661815664403627, "socrates": 0.001082116854881888}, {"ibn_sina": 0.01451873777576505, "einstein": 0.4516646933070591, "darwin": 0.008272527421402968, "sagan": 0.5118031485257966, "hypatia": 0.012812741116550615, "socrates": 0.0009281518534256904}, {"einstein": 0.10362903625195341, "sagan": 0.86767529733103, "socrates": 0.005492137715564503, "ibn_sina": 0.010260664084764424, "darwin": 0.010922420063049217, "hypatia": 0.002020444553638471}, {"socrates": 0.006857877339061137, "ibn_sina": 0.0036707572211138757, "hypatia": 0.007300172427368934, "einstein": 0.5799245263340302, "sagan": 0.39857590945731186, "darwin": 0.0036707572211138757}], "prompt_tokens": 409, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.014556269484544377, "einstein": 0.48203744281320826, "hypatia": 0.011336434073161554, "ibn_sina": 0.008293912344448233, "sagan": 0.48203744281320826, "socrates": 0.0017384984714294077}, "confidence": 0.5163049815451398, "choice": "einstein"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 214.8482580014388, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "d1ab3fa14e7acc3c64152e9b9fefb12997ba7cb18617a7d4ba7df9edddd170aa", "compiled_ids_sha256": "b4fb76f1da222baeff1f762980a440de6c084e5a94b0a0af875b6114f77bd052", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 19.5, 15.75, 15.4375, 19.5, 13.875], "runtime_probabilities": [0.01455626916140318, 0.48203742504119873, 0.011336433701217175, 0.008293911814689636, 0.48203742504119873, 0.0017384984530508518], "decide_seconds": 0.20941857900106697, "probabilities_f64_minus_runtime_max_abs": 1.7772009530681032e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.010178451945149157, "hypatia": 0.002573515675729548, "sagan": 0.7135645706656287, "socrates": 0.002739493176752355, "darwin": 0.008438233040528586, "einstein": 0.26250573549621165}, "confidence": 0.603515074577534, "choice": "sagan"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 213.95201400082442, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "a67d01f47b4e062017ef0625a80f0841b27f2ee07b838a1444e1364e594abbec", "compiled_ids_sha256": "a51e810d1d269d4c0966ebc80dc8ce4fdabfcb95cdc97aeefa02e27a1f643428", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.75, 14.375, 20.0, 14.4375, 15.5625, 19.0], "runtime_probabilities": [0.010178451426327229, 0.0025735157541930676, 0.7135645747184753, 0.0027394932694733143, 0.008438232354819775, 0.26250576972961426], "decide_seconds": 0.20901773299920023, "probabilities_f64_minus_runtime_max_abs": 3.423340261266006e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.004555882128576662, "darwin": 0.002438587976306082, "sagan": 0.21951469405101837, "ibn_sina": 0.006227152548854217, "einstein": 0.7661815664403627, "socrates": 0.001082116854881888}, "confidence": 0.6566671900856555, "choice": "einstein"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 214.2171720006445, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "ffcbf6fb24313235b7d591c15b4afb758fa9a761ccf6d0770a2a4294ee2bcaf7", "compiled_ids_sha256": "6095298403e624d418b700421225f1067cfe6c176c1249a786e99ae584fc212e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.75, 14.125, 18.625, 15.0625, 19.875, 13.3125], "runtime_probabilities": [0.004555882420390844, 0.002438588300719857, 0.21951472759246826, 0.006227153819054365, 0.7661816477775574, 0.0010821169707924128], "decide_seconds": 0.2092864410005859, "probabilities_f64_minus_runtime_max_abs": 8.13371946462027e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.01451873777576505, "einstein": 0.4516646933070591, "darwin": 0.008272527421402968, "sagan": 0.5118031485257966, "hypatia": 0.012812741116550615, "socrates": 0.0009281518534256904}, "confidence": 0.517107945953549, "choice": "sagan"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 214.16713199869264, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "e061af839a84e5f40968432bbf43b1b0645d1b6f93e35f31dfdee2362b65e1b5", "compiled_ids_sha256": "cec607b458d7c156c4c19333a064f264e5ba4ca458982a0cac65ad5d9e6ec02a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 19.375, 15.375, 19.5, 15.8125, 13.1875], "runtime_probabilities": [0.014518738724291325, 0.4516647160053253, 0.00827252771705389, 0.511803150177002, 0.012812742032110691, 0.0009281518869102001], "decide_seconds": 0.20923219100222923, "probabilities_f64_minus_runtime_max_abs": 2.2698266211040874e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.10362903625195341, "sagan": 0.86767529733103, "socrates": 0.005492137715564503, "ibn_sina": 0.010260664084764424, "darwin": 0.010922420063049217, "hypatia": 0.002020444553638471}, "confidence": 0.7234452485322123, "choice": "sagan"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 213.5485670005437, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "067e4724717953ff29f69852ea64f6abec0a34d07cfeaa8a1edce8084c915f87", "compiled_ids_sha256": "d256dd8fd200dd22d074739a7cdb22e47a72af6177a7d756491befc1505fa8b9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 20.25, 15.1875, 15.8125, 15.875, 14.1875], "runtime_probabilities": [0.10362904518842697, 0.8676753640174866, 0.005492137745022774, 0.01026066392660141, 0.010922420769929886, 0.0020204444881528616], "decide_seconds": 0.2085652960013249, "probabilities_f64_minus_runtime_max_abs": 6.668645657903483e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.006857877339061137, "ibn_sina": 0.0036707572211138757, "hypatia": 0.007300172427368934, "einstein": 0.5799245263340302, "sagan": 0.39857590945731186, "darwin": 0.0036707572211138757}, "confidence": 0.5569384214091356, "choice": "einstein"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 214.07868299866095, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "27dec1ce00447b2c51aa4dbbfa98f2d68b019252e4fcb2ce695c4187f86b8845", "compiled_ids_sha256": "16e0789b1b4d0bf6f3bd9dcc79956389be010a3f0f11e87d8bb68082525267a9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.1875, 14.5625, 15.25, 19.625, 19.25, 14.5625], "runtime_probabilities": [0.006857876665890217, 0.0036707569379359484, 0.007300172466784716, 0.5799245238304138, 0.39857590198516846, 0.0036707569379359484], "decide_seconds": 0.20887579400005052, "probabilities_f64_minus_runtime_max_abs": 7.472143404019249e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:37:59Z"} {"id": "c-sagan-01-dessert-03", "uid": "c077", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "prompt_sha256_by_order": ["6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a"], "request_sha256": "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "compiled_prompt_sha256_by_order": ["5323ed669bcbd837510d737e7f7ff27ae1e5464e1a1530c9bd5f04eaf16e3def", "3b2a4e1f58d86cfc7f81af206b22693e4ba145aa2204a6f8a8d20b76e740db2d", "d32f19300ec41a442fabedb83170a6d3e9541c9fdf4b77a2b95e13086fb906d0", "891cf6a7cd0b6302084f885f9deeadf316145175766fdbedb07875afee58a437", "886564d5a6454a34e2cf4fec815135d2c7987a284381791ba4d8b418d1d15f9f", "75029d7e926a01073aaacf9913cc9a5027c779c262e4160a604938a82b96aa5b"], "compiled_tokens_by_order": [331, 331, 331, 331, 331, 331], "choice": "sagan", "probs": {"darwin": 0.006207552942040617, "einstein": 0.0048344470922184355, "hypatia": 0.005831456321859179, "ibn_sina": 0.001474423059371908, "sagan": 0.9807001615088952, "socrates": 0.000951959075614777}, "confidence": 0.9315357721991671, "seconds": 0.18502264450035, "seconds_all_orders": 1.1101358670021, "seconds_by_order": [0.18535423600042122, 0.18492424100259086, 0.18490535099772387, 0.184845759999007, 0.1848048009997001, 0.18530147800265695], "argmax_stable": true, "spread": {"darwin": 0.004534901253536405, "einstein": 0.012226114469494998, "hypatia": 0.004986774497076469, "ibn_sina": 0.0014096448839377444, "sagan": 0.010082800812023951, "socrates": 0.0009773883233814271}, "spread_2dp": {"darwin": 0.01, "einstein": 0.02, "hypatia": 0.01, "ibn_sina": 0.0, "sagan": 0.010000000000000009, "socrates": 0.0}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.006207552942040617, "einstein": 0.0048344470922184355, "hypatia": 0.005831456321859179, "ibn_sina": 0.001474423059371908, "sagan": 0.9807001615088952, "socrates": 0.000951959075614777}, {"ibn_sina": 0.0011545431663329245, "hypatia": 0.0008446818247827098, "sagan": 0.9860477079213223, "socrates": 0.0012290048031428785, "darwin": 0.004860808289267986, "einstein": 0.005863253995151293}, {"hypatia": 0.0021477272813050755, "darwin": 0.0016726516885042122, "sagan": 0.9818221721920221, "ibn_sina": 0.0021477272813050755, "einstein": 0.010907059112050714, "socrates": 0.0013026624448127882}, {"ibn_sina": 0.0019027382374311398, "einstein": 0.003554783020132439, "darwin": 0.002600734815043037, "sagan": 0.9856428089761671, "hypatia": 0.005505755646651271, "socrates": 0.0007931793045750701}, {"einstein": 0.004031241845091842, "sagan": 0.9864123564999038, "socrates": 0.001680471617357502, "ibn_sina": 0.001680471617357502, "darwin": 0.004291234606658073, "hypatia": 0.001904223813631361}, {"socrates": 0.0017705676279564971, "ibn_sina": 0.0007380823973673311, "hypatia": 0.003107442960908199, "einstein": 0.015780897489627438, "sagan": 0.9763295556878798, "darwin": 0.0022734538362606674}], "prompt_tokens": 331, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.006207552942040617, "einstein": 0.0048344470922184355, "hypatia": 0.005831456321859179, "ibn_sina": 0.001474423059371908, "sagan": 0.9807001615088952, "socrates": 0.000951959075614777}, "confidence": 0.9315357721991671, "choice": "sagan"}}, "usage": {"input_tokens": 331, "output_tokens": 1}, "latency_ms": 185.32754799889517, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 331, "compiled_prompt_sha256": "5323ed669bcbd837510d737e7f7ff27ae1e5464e1a1530c9bd5f04eaf16e3def", "compiled_ids_sha256": "9fa1f31f937a41e05c35b71c72dada3a715a34d8f8cc835e0cc5cabd5ccf3784", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.6875, 15.4375, 15.625, 14.25, 20.75, 13.8125], "runtime_probabilities": [0.0062075527384877205, 0.004834447056055069, 0.005831455811858177, 0.0014744229847565293, 0.9807001352310181, 0.000951958994846791], "decide_seconds": 0.181321117997868, "probabilities_f64_minus_runtime_max_abs": 2.6277877096525515e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0011545431663329245, "hypatia": 0.0008446818247827098, "sagan": 0.9860477079213223, "socrates": 0.0012290048031428785, "darwin": 0.004860808289267986, "einstein": 0.005863253995151293}, "confidence": 0.9487093753621608, "choice": "sagan"}}, "usage": {"input_tokens": 331, "output_tokens": 1}, "latency_ms": 184.90150100114988, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 331, "compiled_prompt_sha256": "3b2a4e1f58d86cfc7f81af206b22693e4ba145aa2204a6f8a8d20b76e740db2d", "compiled_ids_sha256": "a39d341f450e8a28df8e350fd5dae9f3ae4bedb1db746cb33100525299d1b4f9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.0, 13.6875, 20.75, 14.0625, 15.4375, 15.625], "runtime_probabilities": [0.0011545431334525347, 0.0008446818683296442, 0.9860477447509766, 0.0012290049344301224, 0.004860808607190847, 0.0058632539585232735], "decide_seconds": 0.181019360999926, "probabilities_f64_minus_runtime_max_abs": 3.6829654215253527e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0021477272813050755, "darwin": 0.0016726516885042122, "sagan": 0.9818221721920221, "ibn_sina": 0.0021477272813050755, "einstein": 0.010907059112050714, "socrates": 0.0013026624448127882}, "confidence": 0.9369168715113225, "choice": "sagan"}}, "usage": {"input_tokens": 331, "output_tokens": 1}, "latency_ms": 184.8829009977635, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 331, "compiled_prompt_sha256": "d32f19300ec41a442fabedb83170a6d3e9541c9fdf4b77a2b95e13086fb906d0", "compiled_ids_sha256": "37ff1b1589d7ddff42626f8c3eb41433e939c949a7974cc697d4218162ac5d5a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.625, 14.375, 20.75, 14.625, 16.25, 14.125], "runtime_probabilities": [0.0021477271802723408, 0.0016726518515497446, 0.98182213306427, 0.0021477271802723408, 0.010907059535384178, 0.0013026624219492078], "decide_seconds": 0.1809799610018672, "probabilities_f64_minus_runtime_max_abs": 3.9127752038403685e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0019027382374311398, "einstein": 0.003554783020132439, "darwin": 0.002600734815043037, "sagan": 0.9856428089761671, "hypatia": 0.005505755646651271, "socrates": 0.0007931793045750701}, "confidence": 0.9464194855606867, "choice": "sagan"}}, "usage": {"input_tokens": 331, "output_tokens": 1}, "latency_ms": 184.8229610004637, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 331, "compiled_prompt_sha256": "891cf6a7cd0b6302084f885f9deeadf316145175766fdbedb07875afee58a437", "compiled_ids_sha256": "0f6135a6ba8e2dd16425ebfaf1e9d945e5ccb52a05e28902a3e4ecc79b5deb93", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.75, 15.375, 15.0625, 21.0, 15.8125, 13.875], "runtime_probabilities": [0.001902738120406866, 0.0035547828301787376, 0.0026007345877587795, 0.9856427311897278, 0.005505755543708801, 0.0007931792642921209], "decide_seconds": 0.18085494200204266, "probabilities_f64_minus_runtime_max_abs": 7.778643928979534e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.004031241845091842, "sagan": 0.9864123564999038, "socrates": 0.001680471617357502, "ibn_sina": 0.001680471617357502, "darwin": 0.004291234606658073, "hypatia": 0.001904223813631361}, "confidence": 0.9483671596482413, "choice": "sagan"}}, "usage": {"input_tokens": 331, "output_tokens": 1}, "latency_ms": 184.78172100003576, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 331, "compiled_prompt_sha256": "886564d5a6454a34e2cf4fec815135d2c7987a284381791ba4d8b418d1d15f9f", "compiled_ids_sha256": "1eb2aa9e51b7d6bf68f29de59d25eb9c740ee1bc861706b27b1c403c0f167a22", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5, 21.0, 14.625, 14.625, 15.5625, 14.75], "runtime_probabilities": [0.004031241871416569, 0.9864122867584229, 0.0016804717015475035, 0.0016804717015475035, 0.004291234537959099, 0.0019042236963286996], "decide_seconds": 0.18093870099983178, "probabilities_f64_minus_runtime_max_abs": 6.974148092364629e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0017705676279564971, "ibn_sina": 0.0007380823973673311, "hypatia": 0.003107442960908199, "einstein": 0.015780897489627438, "sagan": 0.9763295556878798, "darwin": 0.0022734538362606674}, "confidence": 0.9234363870610149, "choice": "sagan"}}, "usage": {"input_tokens": 331, "output_tokens": 1}, "latency_ms": 185.27897800231585, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 331, "compiled_prompt_sha256": "75029d7e926a01073aaacf9913cc9a5027c779c262e4160a604938a82b96aa5b", "compiled_ids_sha256": "acf11c529b163e839cd71470844a2035a26f5f1e36b3a85c914e999481b03def", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.0625, 13.1875, 14.625, 16.25, 20.375, 14.3125], "runtime_probabilities": [0.0017705676145851612, 0.000738082395400852, 0.003107443219050765, 0.015780897811055183, 0.9763295650482178, 0.002273453865200281], "decide_seconds": 0.18121905899897683, "probabilities_f64_minus_runtime_max_abs": 9.360337949182451e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:00Z"} {"id": "c-sagan-01-dessert-03", "uid": "c078", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "prompt_sha256_by_order": ["2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541"], "request_sha256": "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "compiled_prompt_sha256_by_order": ["9185a3cc0cf1abb59a3142933ef89d11e19cf783575408d4218a74be6254ce2d", "629419de84391ebb9e415b1e7674cd3c9f7fc69557a4d8bc4a2fbb9fee405b78", "9f5c19734b49d983625517cb775de8de6568893570aff88022d6f850013da9c2", "3511fd6fe05db9be3342e71a4e01edf6f2892e190e7075a3448d8b10469389f2", "0e23ad47c41478ec22e152aa0e34d025686f509c4fdc87d87f7e2cf6a3be4dcc", "7f109023050a96095c010c7db76a30b2ae9d06086909fda1e15a4b54c3c33587"], "compiled_tokens_by_order": [360, 360, 360, 360, 360, 360], "choice": "sagan", "probs": {"darwin": 0.008822435062436181, "einstein": 0.02552856231438407, "hypatia": 0.005026871970880024, "ibn_sina": 0.0018492828514877304, "sagan": 0.9579522335098852, "socrates": 0.0008206142909267494}, "confidence": 0.8768817726245011, "seconds": 0.1927208429994304, "seconds_all_orders": 1.1563250579965825, "seconds_by_order": [0.19309183700170252, 0.19257458100037184, 0.19268543999714893, 0.19258274099775008, 0.19279642900073668, 0.19259402999887243], "argmax_stable": true, "spread": {"darwin": 0.061479285063297896, "einstein": 0.4451519929419726, "hypatia": 0.004634806337045772, "ibn_sina": 0.00439939154179182, "sagan": 0.44170225909142213, "socrates": 0.009887423546274145}, "spread_2dp": {"darwin": 0.06, "einstein": 0.44999999999999996, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.43999999999999995, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.008822435062436181, "einstein": 0.02552856231438407, "hypatia": 0.005026871970880024, "ibn_sina": 0.0018492828514877304, "sagan": 0.9579522335098852, "socrates": 0.0008206142909267494}, {"ibn_sina": 0.0008835335447162825, "hypatia": 0.0003920656338342519, "sagan": 0.9689121817197086, "socrates": 0.0007797156165417074, "darwin": 0.008923372695844636, "einstein": 0.020109130789354397}, {"hypatia": 0.0016230156534787937, "darwin": 0.0011874236181101323, "sagan": 0.952686542782389, "ibn_sina": 0.002083993350748835, "einstein": 0.04185812539414332, "socrates": 0.0005608992011299464}, {"ibn_sina": 0.0025317523858592516, "einstein": 0.03283226449753015, "darwin": 0.0026950363860992438, "sagan": 0.9594994148563861, "hypatia": 0.00209889644790002, "socrates": 0.0003426354262252318}, {"einstein": 0.055303176306513754, "sagan": 0.865087228989152, "socrates": 0.010230058972499376, "ibn_sina": 0.005143996695877644, "darwin": 0.06266670868140803, "hypatia": 0.0015688303545493049}, {"socrates": 0.0019014158718928832, "ibn_sina": 0.0007446051540858248, "hypatia": 0.002441466307203945, "einstein": 0.465261123731327, "sagan": 0.5272099226282865, "darwin": 0.002441466307203945}], "prompt_tokens": 360, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.008822435062436181, "einstein": 0.02552856231438407, "hypatia": 0.005026871970880024, "ibn_sina": 0.0018492828514877304, "sagan": 0.9579522335098852, "socrates": 0.0008206142909267494}, "confidence": 0.8768817726245011, "choice": "sagan"}}, "usage": {"input_tokens": 360, "output_tokens": 1}, "latency_ms": 193.0644870008109, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 360, "compiled_prompt_sha256": "9185a3cc0cf1abb59a3142933ef89d11e19cf783575408d4218a74be6254ce2d", "compiled_ids_sha256": "36b52d662b922ffae58003f7a5d011a210ce57c6309ab056b16d4638019782b0", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.6875, 16.75, 15.125, 14.125, 20.375, 13.3125], "runtime_probabilities": [0.008822435513138771, 0.0255285631865263, 0.005026872269809246, 0.0018492829985916615, 0.9579522609710693, 0.0008206143393181264], "decide_seconds": 0.1887528699990071, "probabilities_f64_minus_runtime_max_abs": 2.74611841044603e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0008835335447162825, "hypatia": 0.0003920656338342519, "sagan": 0.9689121817197086, "socrates": 0.0007797156165417074, "darwin": 0.008923372695844636, "einstein": 0.020109130789354397}, "confidence": 0.9072778939040007, "choice": "sagan"}}, "usage": {"input_tokens": 360, "output_tokens": 1}, "latency_ms": 192.55058099952294, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 360, "compiled_prompt_sha256": "629419de84391ebb9e415b1e7674cd3c9f7fc69557a4d8bc4a2fbb9fee405b78", "compiled_ids_sha256": "83449f2780bb31841eccaca7e90f626abb698216497b6570f7e86111392cea54", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [13.5, 12.6875, 20.5, 13.375, 15.8125, 16.625], "runtime_probabilities": [0.0008835335611365736, 0.00039206561632454395, 0.9689121842384338, 0.0007797155994921923, 0.008923372253775597, 0.020109130069613457], "decide_seconds": 0.18835814299745834, "probabilities_f64_minus_runtime_max_abs": 2.5187252283842554e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0016230156534787937, "darwin": 0.0011874236181101323, "sagan": 0.952686542782389, "ibn_sina": 0.002083993350748835, "einstein": 0.04185812539414332, "socrates": 0.0005608992011299464}, "confidence": 0.8802854074026142, "choice": "sagan"}}, "usage": {"input_tokens": 360, "output_tokens": 1}, "latency_ms": 192.66140000036103, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 360, "compiled_prompt_sha256": "9f5c19734b49d983625517cb775de8de6568893570aff88022d6f850013da9c2", "compiled_ids_sha256": "aa0d88f3004823454d8f6f25556abab2cd36241475eb4b7b31805101fc8aff49", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.0, 13.6875, 20.375, 14.25, 17.25, 12.9375], "runtime_probabilities": [0.0016230158507823944, 0.0011874237097799778, 0.9526865482330322, 0.0020839935168623924, 0.04185812547802925, 0.0005608992069028318], "decide_seconds": 0.18838049300029525, "probabilities_f64_minus_runtime_max_abs": 5.45064327006628e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0025317523858592516, "einstein": 0.03283226449753015, "darwin": 0.0026950363860992438, "sagan": 0.9594994148563861, "hypatia": 0.00209889644790002, "socrates": 0.0003426354262252318}, "confidence": 0.8891628606043089, "choice": "sagan"}}, "usage": {"input_tokens": 360, "output_tokens": 1}, "latency_ms": 192.55830100155436, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 360, "compiled_prompt_sha256": "3511fd6fe05db9be3342e71a4e01edf6f2892e190e7075a3448d8b10469389f2", "compiled_ids_sha256": "a041de6bb22a3abcce3cd5cb9274a1b041ff0956442c449457c627c257a87c8e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.6875, 17.25, 14.75, 20.625, 14.5, 12.6875], "runtime_probabilities": [0.0025317524559795856, 0.03283226490020752, 0.0026950365863740444, 0.9594994187355042, 0.0020988965407013893, 0.0003426354087423533], "decide_seconds": 0.18829067399929045, "probabilities_f64_minus_runtime_max_abs": 3.87911802679497e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.055303176306513754, "sagan": 0.865087228989152, "socrates": 0.010230058972499376, "ibn_sina": 0.005143996695877644, "darwin": 0.06266670868140803, "hypatia": 0.0015688303545493049}, "confidence": 0.696850553769878, "choice": "sagan"}}, "usage": {"input_tokens": 360, "output_tokens": 1}, "latency_ms": 192.77237899950705, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 360, "compiled_prompt_sha256": "0e23ad47c41478ec22e152aa0e34d025686f509c4fdc87d87f7e2cf6a3be4dcc", "compiled_ids_sha256": "9a513e31a566b29634c7a06654ed5f0715fd0ce6ca53c259a9e8fb5538f73553", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 20.25, 15.8125, 15.125, 17.625, 13.9375], "runtime_probabilities": [0.0553031750023365, 0.8650872111320496, 0.010230057872831821, 0.005143996328115463, 0.06266670674085617, 0.0015688303392380476], "decide_seconds": 0.1886129010017612, "probabilities_f64_minus_runtime_max_abs": 1.7857102463381125e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0019014158718928832, "ibn_sina": 0.0007446051540858248, "hypatia": 0.002441466307203945, "einstein": 0.465261123731327, "sagan": 0.5272099226282865, "darwin": 0.002441466307203945}, "confidence": 0.5869189887480024, "choice": "sagan"}}, "usage": {"input_tokens": 360, "output_tokens": 1}, "latency_ms": 192.57006100087892, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 360, "compiled_prompt_sha256": "7f109023050a96095c010c7db76a30b2ae9d06086909fda1e15a4b54c3c33587", "compiled_ids_sha256": "83290ee5bf70a6206a37c4f9a8deb40adc21e5935919d4b8275ac78cbe29490e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.0, 13.0625, 14.25, 19.5, 19.625, 14.25], "runtime_probabilities": [0.0019014158751815557, 0.000744605204090476, 0.0024414665531367064, 0.4652611315250397, 0.5272099375724792, 0.0024414665531367064], "decide_seconds": 0.18821743499938748, "probabilities_f64_minus_runtime_max_abs": 1.4944192772148313e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:01Z"} {"id": "c-sagan-01-dessert-03", "uid": "c079", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "prompt_sha256_by_order": ["fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520"], "request_sha256": "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "compiled_prompt_sha256_by_order": ["4c9fb986381beb2b96309a2afbd1b2af9fe10613bf4378bdb77d7479f8c41680", "4633d390e8d44b5e0602c3870a0f99da84a4ae306173ae0a3145e5d367fe328c", "5d26fe24f8b07848c7fa8024319fe88766db722129d7d7bffc5883e3ab2aed19", "81e4846b544328a97f2990eae522b97b7e247176c221d1e89c48ba2f68f56510", "be489bc7e710f529e7158e9eb3d0196ec5bec446b9164b7051a0bd158fda5456", "288eaf677a09cb6634d10a4c1ac5a8676a716d57b3249972b2fee3ee7243bf40"], "compiled_tokens_by_order": [397, 397, 397, 397, 397, 397], "choice": "sagan", "probs": {"darwin": 0.04852206369637403, "einstein": 0.03334869416333052, "hypatia": 0.03778902120297668, "ibn_sina": 0.004513254970165509, "sagan": 0.8600741582668064, "socrates": 0.015752807700346833}, "confidence": 0.6632263576196678, "seconds": 0.21230315049918622, "seconds_all_orders": 1.2738189029951172, "seconds_by_order": [0.21192741000049864, 0.2122509379987605, 0.21305518100052723, 0.21230019699942204, 0.21222013799706474, 0.2120650389988441], "argmax_stable": true, "spread": {"darwin": 0.12951384692473653, "einstein": 0.05290680561998965, "hypatia": 0.0914034246125384, "ibn_sina": 0.025953692148793287, "sagan": 0.3073367698611671, "socrates": 0.03208959971598038}, "spread_2dp": {"darwin": 0.13, "einstein": 0.05, "hypatia": 0.1, "ibn_sina": 0.03, "sagan": 0.30000000000000004, "socrates": 0.04}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.04852206369637403, "einstein": 0.03334869416333052, "hypatia": 0.03778902120297668, "ibn_sina": 0.004513254970165509, "sagan": 0.8600741582668064, "socrates": 0.015752807700346833}, {"ibn_sina": 0.0128940912915643, "hypatia": 0.0047434710987542365, "sagan": 0.9039455868309523, "socrates": 0.014610919600738523, "darwin": 0.03971659724776362, "einstein": 0.02408933393022703}, {"hypatia": 0.013326977967912776, "darwin": 0.01037906087738583, "sagan": 0.9342932857757494, "ibn_sina": 0.006701225361189545, "einstein": 0.021972472049849763, "socrates": 0.013326977967912776}, {"ibn_sina": 0.030466947118958795, "einstein": 0.0268869864636898, "darwin": 0.05691975079029606, "sagan": 0.7857529224370933, "hypatia": 0.07308640672627231, "socrates": 0.0268869864636898}, {"einstein": 0.07487927766983941, "sagan": 0.6269565159145823, "socrates": 0.04541657768389316, "ibn_sina": 0.016707825218270012, "darwin": 0.13989290780212235, "hypatia": 0.09614689571129265}, {"socrates": 0.03608302298377248, "ibn_sina": 0.00627028919550319, "hypatia": 0.040887421676036526, "einstein": 0.0674119618213669, "sagan": 0.8212458177679753, "darwin": 0.028101486555345504}], "prompt_tokens": 397, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.04852206369637403, "einstein": 0.03334869416333052, "hypatia": 0.03778902120297668, "ibn_sina": 0.004513254970165509, "sagan": 0.8600741582668064, "socrates": 0.015752807700346833}, "confidence": 0.6632263576196678, "choice": "sagan"}}, "usage": {"input_tokens": 397, "output_tokens": 1}, "latency_ms": 211.89914099886664, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 397, "compiled_prompt_sha256": "4c9fb986381beb2b96309a2afbd1b2af9fe10613bf4378bdb77d7479f8c41680", "compiled_ids_sha256": "86437e7858dab9217d982273312aa4f196d2a66f6de0c27248ed494a7c977df0", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 17.0, 17.125, 15.0, 20.25, 16.25], "runtime_probabilities": [0.048522062599658966, 0.0333486907184124, 0.03778902068734169, 0.004513254854828119, 0.8600741028785706, 0.01575280725955963], "decide_seconds": 0.20729838599800132, "probabilities_f64_minus_runtime_max_abs": 5.538823588668862e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0128940912915643, "hypatia": 0.0047434710987542365, "sagan": 0.9039455868309523, "socrates": 0.014610919600738523, "darwin": 0.03971659724776362, "einstein": 0.02408933393022703}, "confidence": 0.7475121654235646, "choice": "sagan"}}, "usage": {"input_tokens": 397, "output_tokens": 1}, "latency_ms": 212.22561800095718, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 397, "compiled_prompt_sha256": "4633d390e8d44b5e0602c3870a0f99da84a4ae306173ae0a3145e5d367fe328c", "compiled_ids_sha256": "29bbb1050c1a409fbef1f2270dbc3cd5e252f9a1569e55ceb84c7330bfb392fa", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 15.25, 20.5, 16.375, 17.375, 16.875], "runtime_probabilities": [0.012894091196358204, 0.004743471276015043, 0.9039456248283386, 0.01461092010140419, 0.03971659764647484, 0.024089334532618523], "decide_seconds": 0.20768024300195975, "probabilities_f64_minus_runtime_max_abs": 3.7997386348465056e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.013326977967912776, "darwin": 0.01037906087738583, "sagan": 0.9342932857757494, "ibn_sina": 0.006701225361189545, "einstein": 0.021972472049849763, "socrates": 0.013326977967912776}, "confidence": 0.8083258146200906, "choice": "sagan"}}, "usage": {"input_tokens": 397, "output_tokens": 1}, "latency_ms": 213.02809199914918, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 397, "compiled_prompt_sha256": "5d26fe24f8b07848c7fa8024319fe88766db722129d7d7bffc5883e3ab2aed19", "compiled_ids_sha256": "4813f1dcd6bb58eb899d39cd9b012a9ff146d33efc73347b7b9bed3667ce65a9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 16.25, 20.75, 15.8125, 17.0, 16.5], "runtime_probabilities": [0.01332697831094265, 0.010379062034189701, 0.9342933297157288, 0.0067012254148721695, 0.021972473710775375, 0.01332697831094265], "decide_seconds": 0.20797852000032435, "probabilities_f64_minus_runtime_max_abs": 4.393997932172766e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.030466947118958795, "einstein": 0.0268869864636898, "darwin": 0.05691975079029606, "sagan": 0.7857529224370933, "hypatia": 0.07308640672627231, "socrates": 0.0268869864636898}, "confidence": 0.528612965586907, "choice": "sagan"}}, "usage": {"input_tokens": 397, "output_tokens": 1}, "latency_ms": 212.27431800070917, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 397, "compiled_prompt_sha256": "81e4846b544328a97f2990eae522b97b7e247176c221d1e89c48ba2f68f56510", "compiled_ids_sha256": "daef1278477a90357311568da60bcdc7305a2eb22a12173c7047a73b5841b367", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 16.625, 17.375, 20.0, 17.625, 16.625], "runtime_probabilities": [0.030466943979263306, 0.026886984705924988, 0.056919749826192856, 0.7857528924942017, 0.07308639585971832, 0.026886984705924988], "decide_seconds": 0.20743498500087298, "probabilities_f64_minus_runtime_max_abs": 2.994289161062369e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.07487927766983941, "sagan": 0.6269565159145823, "socrates": 0.04541657768389316, "ibn_sina": 0.016707825218270012, "darwin": 0.13989290780212235, "hypatia": 0.09614689571129265}, "confidence": 0.3325578834951076, "choice": "sagan"}}, "usage": {"input_tokens": 397, "output_tokens": 1}, "latency_ms": 212.18942799896467, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 397, "compiled_prompt_sha256": "be489bc7e710f529e7158e9eb3d0196ec5bec446b9164b7051a0bd158fda5456", "compiled_ids_sha256": "dbf17730bb89eba96c246ae04ed042811507045da046d7835c92daec8dbba1a6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 19.75, 17.125, 16.125, 18.25, 17.875], "runtime_probabilities": [0.07487927377223969, 0.6269564628601074, 0.04541657492518425, 0.016707824543118477, 0.13989290595054626, 0.09614688903093338], "decide_seconds": 0.20736062600190053, "probabilities_f64_minus_runtime_max_abs": 5.305447492442994e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.03608302298377248, "ibn_sina": 0.00627028919550319, "hypatia": 0.040887421676036526, "einstein": 0.0674119618213669, "sagan": 0.8212458177679753, "darwin": 0.028101486555345504}, "confidence": 0.594647268482221, "choice": "sagan"}}, "usage": {"input_tokens": 397, "output_tokens": 1}, "latency_ms": 212.04019899960258, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 397, "compiled_prompt_sha256": "288eaf677a09cb6634d10a4c1ac5a8676a716d57b3249972b2fee3ee7243bf40", "compiled_ids_sha256": "abeb2e23820f04985eda3cf3420b0e1d784fd5cdfed661960f8f810b7d351c10", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 15.25, 17.125, 17.625, 20.125, 16.75], "runtime_probabilities": [0.036083023995161057, 0.006270288955420256, 0.04088741913437843, 0.06741195917129517, 0.8212457895278931, 0.02810148522257805], "decide_seconds": 0.20749506400170503, "probabilities_f64_minus_runtime_max_abs": 2.824008227264585e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:02Z"} {"id": "c-sagan-01-main-01", "uid": "c080", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "prompt_sha256_by_order": ["140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631"], "request_sha256": "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "compiled_prompt_sha256_by_order": ["9039ec9ebd536e8ff815ec4d3af1f3487d4b1b3ac68f56b8c74d168136f644a8", "d6f6d0e89e0ff3c08e90886efdbf4fc85b17286e73cab77a600bf5a29c8c244a", "634404df4f0f7dd3981cb1ed7a904e2c0c257d58f839af054cc97539723d3ee5", "bc6965f7530f96e3349a0d9519fd5cc3804aa52c359cc414420aafe2f7456b0a", "e9f387a197c782c1116fc9818b4fcee6083b269a08ed87d757dccea28b386fb2", "acec5b43c52adb11448d98ec15993c4c089d7c3bbe6449e7af85b598b61ebfd3"], "compiled_tokens_by_order": [333, 333, 333, 333, 333, 333], "choice": "sagan", "probs": {"darwin": 0.009765663532996006, "einstein": 0.10499077741402507, "hypatia": 0.0027979094467259873, "ibn_sina": 0.0020469942796021838, "sagan": 0.8790770164864193, "socrates": 0.0013216388402314742}, "confidence": 0.758325714598002, "seconds": 0.18587186616665954, "seconds_all_orders": 1.1152311969999573, "seconds_by_order": [0.1857848039981036, 0.1873846410016995, 0.18575263299862854, 0.1853556070018385, 0.18527557700144826, 0.18567793499823892], "argmax_stable": true, "spread": {"darwin": 0.016465780197751842, "einstein": 0.06338638565869689, "hypatia": 0.008978609702060932, "ibn_sina": 0.009455446105333946, "sagan": 0.06806738190646033, "socrates": 0.015880478700952508}, "spread_2dp": {"darwin": 0.02, "einstein": 0.060000000000000005, "hypatia": 0.01, "ibn_sina": 0.01, "sagan": 0.06999999999999995, "socrates": 0.02}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.009765663532996006, "einstein": 0.10499077741402507, "hypatia": 0.0027979094467259873, "ibn_sina": 0.0020469942796021838, "sagan": 0.8790770164864193, "socrates": 0.0013216388402314742}, {"ibn_sina": 0.011502440384936129, "hypatia": 0.0037342956516118403, "sagan": 0.9137520186387772, "socrates": 0.006553892973827954, "darwin": 0.018964318127604365, "einstein": 0.04549303422324249}, {"hypatia": 0.008955565757686012, "darwin": 0.004503141261559783, "sagan": 0.9134925418084417, "ibn_sina": 0.010147985484660145, "einstein": 0.05839762442609258, "socrates": 0.004503141261559783}, {"ibn_sina": 0.0056305421420943325, "einstein": 0.04160439175532818, "darwin": 0.0024985379298525217, "sagan": 0.9469115917812577, "hypatia": 0.0028312143901411753, "socrates": 0.0005237220013260195}, {"einstein": 0.05052849539619133, "sagan": 0.8956390111168132, "socrates": 0.016404200702278528, "ibn_sina": 0.008248550124478731, "darwin": 0.01858839464958466, "hypatia": 0.010591348010653524}, {"socrates": 0.015121349907152292, "ibn_sina": 0.004909183557145944, "hypatia": 0.01177651914878692, "einstein": 0.08174524527725342, "sagan": 0.8788442098747974, "darwin": 0.007603492234864128}], "prompt_tokens": 333, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.009765663532996006, "einstein": 0.10499077741402507, "hypatia": 0.0027979094467259873, "ibn_sina": 0.0020469942796021838, "sagan": 0.8790770164864193, "socrates": 0.0013216388402314742}, "confidence": 0.758325714598002, "choice": "sagan"}}, "usage": {"input_tokens": 333, "output_tokens": 1}, "latency_ms": 185.75791400144226, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 333, "compiled_prompt_sha256": "9039ec9ebd536e8ff815ec4d3af1f3487d4b1b3ac68f56b8c74d168136f644a8", "compiled_ids_sha256": "d5b1938a23c4d5daa57a1a68304eaa2e03f3dfd41e38df0bac9ead15b4c8f182", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.625, 18.0, 14.375, 14.0625, 20.125, 13.625], "runtime_probabilities": [0.009765664115548134, 0.10499077290296555, 0.0027979095466434956, 0.0020469941664487123, 0.8790770173072815, 0.0013216388178989291], "decide_seconds": 0.18162415600090753, "probabilities_f64_minus_runtime_max_abs": 4.511059525924566e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.011502440384936129, "hypatia": 0.0037342956516118403, "sagan": 0.9137520186387772, "socrates": 0.006553892973827954, "darwin": 0.018964318127604365, "einstein": 0.04549303422324249}, "confidence": 0.7748673578076831, "choice": "sagan"}}, "usage": {"input_tokens": 333, "output_tokens": 1}, "latency_ms": 187.35891100004665, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 333, "compiled_prompt_sha256": "d6f6d0e89e0ff3c08e90886efdbf4fc85b17286e73cab77a600bf5a29c8c244a", "compiled_ids_sha256": "de128e35481af2b6834947a1950201ca68fafdef653a21826bc449d4c3b2b234", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.75, 14.625, 20.125, 15.1875, 16.25, 17.125], "runtime_probabilities": [0.011502440087497234, 0.0037342957220971584, 0.913752019405365, 0.006553892511874437, 0.018964316695928574, 0.04549303278326988], "decide_seconds": 0.18288511599894264, "probabilities_f64_minus_runtime_max_abs": 1.4399726064318052e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.008955565757686012, "darwin": 0.004503141261559783, "sagan": 0.9134925418084417, "ibn_sina": 0.010147985484660145, "einstein": 0.05839762442609258, "socrates": 0.004503141261559783}, "confidence": 0.7845665985718984, "choice": "sagan"}}, "usage": {"input_tokens": 333, "output_tokens": 1}, "latency_ms": 185.72822400165023, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 333, "compiled_prompt_sha256": "634404df4f0f7dd3981cb1ed7a904e2c0c257d58f839af054cc97539723d3ee5", "compiled_ids_sha256": "09a03a9a11ea7ea49304c88d259b543b72940095202b57c28b0db96b7a417e26", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5, 14.8125, 20.125, 15.625, 17.375, 14.8125], "runtime_probabilities": [0.008955566212534904, 0.0045031411573290825, 0.9134925007820129, 0.010147986002266407, 0.05839762091636658, 0.0045031411573290825], "decide_seconds": 0.18154097699880367, "probabilities_f64_minus_runtime_max_abs": 4.1026428809765036e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0056305421420943325, "einstein": 0.04160439175532818, "darwin": 0.0024985379298525217, "sagan": 0.9469115917812577, "hypatia": 0.0028312143901411753, "socrates": 0.0005237220013260195}, "confidence": 0.8612318110546935, "choice": "sagan"}}, "usage": {"input_tokens": 333, "output_tokens": 1}, "latency_ms": 185.33229699824005, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 333, "compiled_prompt_sha256": "bc6965f7530f96e3349a0d9519fd5cc3804aa52c359cc414420aafe2f7456b0a", "compiled_ids_sha256": "b1dea77649d622bea84c135bb44656e0a4ac92d2bf541d61138ef152418d50d1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.25, 17.25, 14.4375, 20.375, 14.5625, 12.875], "runtime_probabilities": [0.005630542524158955, 0.04160439595580101, 0.0024985382333397865, 0.946911633014679, 0.0028312145732343197, 0.0005237220320850611], "decide_seconds": 0.18134539899983793, "probabilities_f64_minus_runtime_max_abs": 4.1233421232966805e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.05052849539619133, "sagan": 0.8956390111168132, "socrates": 0.016404200702278528, "ibn_sina": 0.008248550124478731, "darwin": 0.01858839464958466, "hypatia": 0.010591348010653524}, "confidence": 0.7327775928429449, "choice": "sagan"}}, "usage": {"input_tokens": 333, "output_tokens": 1}, "latency_ms": 185.2516269973421, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 333, "compiled_prompt_sha256": "e9f387a197c782c1116fc9818b4fcee6083b269a08ed87d757dccea28b386fb2", "compiled_ids_sha256": "ca7308b553e834c9c5addc2626505d0add7bd63a03330e9ea36806d63f6dd168", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 20.0, 16.0, 15.3125, 16.125, 15.5625], "runtime_probabilities": [0.050528500229120255, 0.8956390023231506, 0.016404202207922935, 0.00824854988604784, 0.018588393926620483, 0.010591347701847553], "decide_seconds": 0.18116589000055683, "probabilities_f64_minus_runtime_max_abs": 8.793662575889982e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.015121349907152292, "ibn_sina": 0.004909183557145944, "hypatia": 0.01177651914878692, "einstein": 0.08174524527725342, "sagan": 0.8788442098747974, "darwin": 0.007603492234864128}, "confidence": 0.7225675101160435, "choice": "sagan"}}, "usage": {"input_tokens": 333, "output_tokens": 1}, "latency_ms": 185.6545949995052, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 333, "compiled_prompt_sha256": "acec5b43c52adb11448d98ec15993c4c089d7c3bbe6449e7af85b598b61ebfd3", "compiled_ids_sha256": "cf87871ca9b9eaca3f396be607278ce4ece6257cf2ffa8793f97646a1fbcdfe9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.8125, 14.6875, 15.5625, 17.5, 19.875, 15.125], "runtime_probabilities": [0.015121350064873695, 0.004909184295684099, 0.011776519939303398, 0.08174524456262589, 0.8788442611694336, 0.007603492587804794], "decide_seconds": 0.18164777600031812, "probabilities_f64_minus_runtime_max_abs": 5.129463620523467e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:04Z"} {"id": "c-sagan-01-main-02", "uid": "c081", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "prompt_sha256_by_order": ["365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3"], "request_sha256": "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "compiled_prompt_sha256_by_order": ["e899284aeb4ab20e1a02f56556620f4fbd0c64a5bfca3086c98d666d6ed9d79b", "84be2b1a1fa4ad7b85a75275ba7652beae8de0a0ad537f6a84c2359c868bc409", "da020ac09be783b39bd67ca45116b423d2f97a5b0a0d1c902937ce9be596e7a9", "caf4e2a091ac5fd1feeadb248dbe1d0d2a062523f10f2d26ea30bad9466c54f8", "791f873348b9933d8fff5f837bbf63f77beb0db4d82fed4393b6e99a8d058dcd", "49981d7b5cb7600ad39ead1b3834a2e527bcbf1f27e0eb00c993b68564a5bb4a"], "compiled_tokens_by_order": [366, 366, 366, 366, 366, 366], "choice": "sagan", "probs": {"darwin": 0.030801156409628692, "einstein": 0.09487424093909791, "hypatia": 0.06520604863087354, "ibn_sina": 0.00939382196131095, "sagan": 0.7943722936469538, "socrates": 0.005352438412135132}, "confidence": 0.5739512049643544, "seconds": 0.19445650583353805, "seconds_all_orders": 1.1667390350012283, "seconds_by_order": [0.19497114199839416, 0.1943212570004107, 0.19410985800277558, 0.1943402770011744, 0.19466490400009206, 0.19433159699838143], "argmax_stable": true, "spread": {"darwin": 0.08662758092296212, "einstein": 0.2067923724203203, "hypatia": 0.06102899242614617, "ibn_sina": 0.035012274847952066, "sagan": 0.2700048814076652, "socrates": 0.051918693436299825}, "spread_2dp": {"darwin": 0.08, "einstein": 0.2, "hypatia": 0.07, "ibn_sina": 0.04, "sagan": 0.27, "socrates": 0.06}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.030801156409628692, "einstein": 0.09487424093909791, "hypatia": 0.06520604863087354, "ibn_sina": 0.00939382196131095, "sagan": 0.7943722936469538, "socrates": 0.005352438412135132}, {"ibn_sina": 0.00392398116282697, "hypatia": 0.004177056204727368, "sagan": 0.9601665329005337, "socrates": 0.004733224776760006, "darwin": 0.009413141465497902, "einstein": 0.01758606348965408}, {"hypatia": 0.01024509335237031, "darwin": 0.007495470416611463, "sagan": 0.9222339134864215, "ibn_sina": 0.007495470416611463, "einstein": 0.045915322901911186, "socrates": 0.006614729426074084}, {"ibn_sina": 0.017745748227750213, "einstein": 0.0425698369965104, "darwin": 0.0425698369965104, "sagan": 0.85503803280749, "hypatia": 0.033153422388104356, "socrates": 0.008923122583634669}, {"einstein": 0.06419503348639712, "sagan": 0.6901616514928685, "socrates": 0.05665191821305983, "ibn_sina": 0.03893625601077903, "darwin": 0.09340322258383574, "hypatia": 0.05665191821305983}, {"socrates": 0.02087044560605386, "ibn_sina": 0.006775641660873614, "hypatia": 0.05006559633844678, "einstein": 0.2243784359099744, "sagan": 0.6911342388237777, "darwin": 0.006775641660873614}], "prompt_tokens": 366, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.030801156409628692, "einstein": 0.09487424093909791, "hypatia": 0.06520604863087354, "ibn_sina": 0.00939382196131095, "sagan": 0.7943722936469538, "socrates": 0.005352438412135132}, "confidence": 0.5739512049643544, "choice": "sagan"}}, "usage": {"input_tokens": 366, "output_tokens": 1}, "latency_ms": 194.94394199864473, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 366, "compiled_prompt_sha256": "e899284aeb4ab20e1a02f56556620f4fbd0c64a5bfca3086c98d666d6ed9d79b", "compiled_ids_sha256": "baf3a4db52592f705e8840880143e8aa8842446449b21f0d838487828fea0bfd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 18.25, 17.875, 15.9375, 20.375, 15.375], "runtime_probabilities": [0.030801156535744667, 0.09487424045801163, 0.06520605087280273, 0.009393821470439434, 0.7943723201751709, 0.005352438427507877], "decide_seconds": 0.19039423700087355, "probabilities_f64_minus_runtime_max_abs": 2.6528217067500748e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.00392398116282697, "hypatia": 0.004177056204727368, "sagan": 0.9601665329005337, "socrates": 0.004733224776760006, "darwin": 0.009413141465497902, "einstein": 0.01758606348965408}, "confidence": 0.875000818462714, "choice": "sagan"}}, "usage": {"input_tokens": 366, "output_tokens": 1}, "latency_ms": 194.29744800072513, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 366, "compiled_prompt_sha256": "84be2b1a1fa4ad7b85a75275ba7652beae8de0a0ad537f6a84c2359c868bc409", "compiled_ids_sha256": "f987da62115aacea72c40420bceff1f947b1c69d6be7e988b2df76c44a519863", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.375, 15.4375, 20.875, 15.5625, 16.25, 16.875], "runtime_probabilities": [0.003923981450498104, 0.004177056718617678, 0.9601666331291199, 0.004733225330710411, 0.009413142688572407, 0.01758606545627117], "decide_seconds": 0.19002145000195014, "probabilities_f64_minus_runtime_max_abs": 1.0022858620839514e-07}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.01024509335237031, "darwin": 0.007495470416611463, "sagan": 0.9222339134864215, "ibn_sina": 0.007495470416611463, "einstein": 0.045915322901911186, "socrates": 0.006614729426074084}, "confidence": 0.7937170642712689, "choice": "sagan"}}, "usage": {"input_tokens": 366, "output_tokens": 1}, "latency_ms": 194.08523899983265, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 366, "compiled_prompt_sha256": "da020ac09be783b39bd67ca45116b423d2f97a5b0a0d1c902937ce9be596e7a9", "compiled_ids_sha256": "a5924d483e2892273033c75267402039f6edffc41769696b4dcb3c95911bd209", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 15.9375, 20.75, 15.9375, 17.75, 15.8125], "runtime_probabilities": [0.010245095007121563, 0.007495471742004156, 0.9222339987754822, 0.007495471742004156, 0.04591532424092293, 0.006614729296416044], "decide_seconds": 0.18979927300097188, "probabilities_f64_minus_runtime_max_abs": 8.528906070282005e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.017745748227750213, "einstein": 0.0425698369965104, "darwin": 0.0425698369965104, "sagan": 0.85503803280749, "hypatia": 0.033153422388104356, "socrates": 0.008923122583634669}, "confidence": 0.6488069381317132, "choice": "sagan"}}, "usage": {"input_tokens": 366, "output_tokens": 1}, "latency_ms": 194.31645700024092, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 366, "compiled_prompt_sha256": "caf4e2a091ac5fd1feeadb248dbe1d0d2a062523f10f2d26ea30bad9466c54f8", "compiled_ids_sha256": "45832ddbd3efe171a06ae499255398d2b2e0b466e7f9f4db14161e02889178ae", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 17.375, 17.375, 20.375, 17.125, 15.8125], "runtime_probabilities": [0.017745748162269592, 0.0425698347389698, 0.0425698347389698, 0.855038046836853, 0.03315342217683792, 0.0089231226593256], "decide_seconds": 0.18986626100013382, "probabilities_f64_minus_runtime_max_abs": 1.402936300465285e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.06419503348639712, "sagan": 0.6901616514928685, "socrates": 0.05665191821305983, "ibn_sina": 0.03893625601077903, "darwin": 0.09340322258383574, "hypatia": 0.05665191821305983}, "confidence": 0.3831200967091969, "choice": "sagan"}}, "usage": {"input_tokens": 366, "output_tokens": 1}, "latency_ms": 194.64023399996222, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 366, "compiled_prompt_sha256": "791f873348b9933d8fff5f837bbf63f77beb0db4d82fed4393b6e99a8d058dcd", "compiled_ids_sha256": "f8442972e84cb078f152f8cae1731c470b3ea5c422e1d6f83df994bdc447c159", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 20.0, 17.5, 17.125, 18.0, 17.5], "runtime_probabilities": [0.06419502943754196, 0.6901615858078003, 0.05665191635489464, 0.03893625736236572, 0.09340322017669678, 0.05665191635489464], "decide_seconds": 0.19031424900094862, "probabilities_f64_minus_runtime_max_abs": 6.568506816240216e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.02087044560605386, "ibn_sina": 0.006775641660873614, "hypatia": 0.05006559633844678, "einstein": 0.2243784359099744, "sagan": 0.6911342388237777, "darwin": 0.006775641660873614}, "confidence": 0.5038450206916103, "choice": "sagan"}}, "usage": {"input_tokens": 366, "output_tokens": 1}, "latency_ms": 194.30812700011302, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 366, "compiled_prompt_sha256": "49981d7b5cb7600ad39ead1b3834a2e527bcbf1f27e0eb00c993b68564a5bb4a", "compiled_ids_sha256": "805e50fd5896bf2ddfaa93d190dd564ed9384895196e475151b3dd24a10df0e4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 15.5, 17.5, 19.0, 20.125, 15.5], "runtime_probabilities": [0.020870445296168327, 0.006775642279535532, 0.050065599381923676, 0.22437843680381775, 0.6911342144012451, 0.006775642279535532], "decide_seconds": 0.19001163099892437, "probabilities_f64_minus_runtime_max_abs": 2.4422532551859888e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:05Z"} {"id": "c-sagan-01-main-02", "uid": "c082", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "prompt_sha256_by_order": ["dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc"], "request_sha256": "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "compiled_prompt_sha256_by_order": ["6b3262c0b05359df640a75374ab1cb9ee093c35d018b01ad37614f989d8ad180", "e58f881f1f8ad60e90fb06493132e99ff4ff5bdcd62b203c274caadf5f8e7bd4", "51144324b9565a106912a27b203efc22e0622c9d2a6c6bf22976aed0d6f4c980", "9861ffca80b3747e09e69887723701d4efc811d7e5b8ba2e3cca712dcc6e7fd9", "a4cdd1d51e4bd6a3833600e0339a5be23818ee80cc31a34409d53b098778e816", "d6853c19ae77d785ff4d0e26d079aa1237f28072b8649be382018b78c39bc252"], "compiled_tokens_by_order": [316, 316, 316, 316, 316, 316], "choice": "einstein", "probs": {"darwin": 0.021926784835390067, "einstein": 0.8227965307281497, "hypatia": 0.06753925209256888, "ibn_sina": 0.00858665231647026, "sagan": 0.07653199902998484, "socrates": 0.002618780997436331}, "confidence": 0.6208323917429652, "seconds": 0.17759746366634013, "seconds_all_orders": 1.0655847819980409, "seconds_by_order": [0.17755783699976746, 0.1776614469999913, 0.17743597800290445, 0.17793976399843814, 0.17752996799754328, 0.1774597879993962], "argmax_stable": true, "spread": {"darwin": 0.050428030491605186, "einstein": 0.32408109777396976, "hypatia": 0.10157111882164023, "ibn_sina": 0.10493862099192913, "sagan": 0.0987822477249346, "socrates": 0.029906754370284207}, "spread_2dp": {"darwin": 0.05, "einstein": 0.31999999999999995, "hypatia": 0.1, "ibn_sina": 0.1, "sagan": 0.09999999999999999, "socrates": 0.03}, "choices_by_order": ["einstein", "einstein", "einstein", "einstein", "einstein", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.021926784835390067, "einstein": 0.8227965307281497, "hypatia": 0.06753925209256888, "ibn_sina": 0.00858665231647026, "sagan": 0.07653199902998484, "socrates": 0.002618780997436331}, {"ibn_sina": 0.051389842430404606, "hypatia": 0.1087922972788893, "sagan": 0.09600886537368232, "socrates": 0.010119259961450543, "darwin": 0.024274842714754153, "einstein": 0.709414892240819}, {"hypatia": 0.1691103709142091, "darwin": 0.017824101959726897, "sagan": 0.1162277448617006, "ibn_sina": 0.022886599945923118, "einstein": 0.6688444916072535, "socrates": 0.005106690711186871}, {"ibn_sina": 0.03676265514346533, "einstein": 0.7383976672784879, "darwin": 0.008202857129642915, "sagan": 0.047204183589135806, "hypatia": 0.164758789753076, "socrates": 0.004673847106192132}, {"einstein": 0.5087849765933953, "sagan": 0.14576933636440728, "socrates": 0.032525535367720536, "ibn_sina": 0.1135252733083994, "darwin": 0.053625542001670155, "hypatia": 0.14576933636440728}, {"socrates": 0.022195129083964198, "ibn_sina": 0.01728558391096223, "hypatia": 0.07746861248817094, "einstein": 0.832866074367365, "sagan": 0.04698708863947267, "darwin": 0.0031975115100649695}], "prompt_tokens": 316, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.021926784835390067, "einstein": 0.8227965307281497, "hypatia": 0.06753925209256888, "ibn_sina": 0.00858665231647026, "sagan": 0.07653199902998484, "socrates": 0.002618780997436331}, "confidence": 0.6208323917429652, "choice": "einstein"}}, "usage": {"input_tokens": 316, "output_tokens": 1}, "latency_ms": 177.53242799881264, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 316, "compiled_prompt_sha256": "6b3262c0b05359df640a75374ab1cb9ee093c35d018b01ad37614f989d8ad180", "compiled_ids_sha256": "9cf18e6fda3d59410ee91dbb91dadbf48ceba6733de0f541fd3d4d6902aa5f4c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 20.125, 17.625, 15.5625, 17.75, 14.375], "runtime_probabilities": [0.02192678488790989, 0.8227965235710144, 0.06753925234079361, 0.00858665257692337, 0.07653199881315231, 0.0026187810581177473], "decide_seconds": 0.17361346899997443, "probabilities_f64_minus_runtime_max_abs": 7.157135328839104e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.051389842430404606, "hypatia": 0.1087922972788893, "sagan": 0.09600886537368232, "socrates": 0.010119259961450543, "darwin": 0.024274842714754153, "einstein": 0.709414892240819}, "confidence": 0.44236331113708205, "choice": "einstein"}}, "usage": {"input_tokens": 316, "output_tokens": 1}, "latency_ms": 177.63881700011552, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 316, "compiled_prompt_sha256": "e58f881f1f8ad60e90fb06493132e99ff4ff5bdcd62b203c274caadf5f8e7bd4", "compiled_ids_sha256": "3fd853485a7de4c9711c7e80a7f456549e61f5bff87ad19f7d61cf288cd8ea90", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.25, 18.0, 17.875, 15.625, 16.5, 19.875], "runtime_probabilities": [0.051389843225479126, 0.10879229754209518, 0.09600885957479477, 0.010119259357452393, 0.02427484095096588, 0.7094148397445679], "decide_seconds": 0.17381503599972348, "probabilities_f64_minus_runtime_max_abs": 5.249625112391243e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.1691103709142091, "darwin": 0.017824101959726897, "sagan": 0.1162277448617006, "ibn_sina": 0.022886599945923118, "einstein": 0.6688444916072535, "socrates": 0.005106690711186871}, "confidence": 0.439166480041872, "choice": "einstein"}}, "usage": {"input_tokens": 316, "output_tokens": 1}, "latency_ms": 177.41262900017318, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 316, "compiled_prompt_sha256": "51144324b9565a106912a27b203efc22e0622c9d2a6c6bf22976aed0d6f4c980", "compiled_ids_sha256": "b70e47b4e50f336958b0e276a601b674909a0444cca8245c459f864bb91e16c3", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.375, 16.125, 18.0, 16.375, 19.75, 14.875], "runtime_probabilities": [0.16911038756370544, 0.01782410219311714, 0.11622774600982666, 0.022886600345373154, 0.6688445210456848, 0.005106690805405378], "decide_seconds": 0.1735832490012399, "probabilities_f64_minus_runtime_max_abs": 2.9438431359274375e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.03676265514346533, "einstein": 0.7383976672784879, "darwin": 0.008202857129642915, "sagan": 0.047204183589135806, "hypatia": 0.164758789753076, "socrates": 0.004673847106192132}, "confidence": 0.525000630631097, "choice": "einstein"}}, "usage": {"input_tokens": 316, "output_tokens": 1}, "latency_ms": 177.9171740017773, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 316, "compiled_prompt_sha256": "9861ffca80b3747e09e69887723701d4efc811d7e5b8ba2e3cca712dcc6e7fd9", "compiled_ids_sha256": "a3c5e21cefdb935f6a684b8afb1807ec59a612ca14af69e48bb301e7359ed68f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 19.75, 15.25, 17.0, 18.25, 14.6875], "runtime_probabilities": [0.036762651056051254, 0.7383975982666016, 0.008202857337892056, 0.04720417782664299, 0.16475877165794373, 0.004673846531659365], "decide_seconds": 0.17386470599740278, "probabilities_f64_minus_runtime_max_abs": 6.901188631669442e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.5087849765933953, "sagan": 0.14576933636440728, "socrates": 0.032525535367720536, "ibn_sina": 0.1135252733083994, "darwin": 0.053625542001670155, "hypatia": 0.14576933636440728}, "confidence": 0.20717944480661676, "choice": "einstein"}}, "usage": {"input_tokens": 316, "output_tokens": 1}, "latency_ms": 177.5075180012209, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 316, "compiled_prompt_sha256": "a4cdd1d51e4bd6a3833600e0339a5be23818ee80cc31a34409d53b098778e816", "compiled_ids_sha256": "529c224f0eda1a82a9fdda9e8ffc63262bb8440329bcbfe19872bd59d1052dce", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.5, 18.25, 16.75, 18.0, 17.25, 18.25], "runtime_probabilities": [0.5087849497795105, 0.14576932787895203, 0.03252553194761276, 0.11352527141571045, 0.05362553521990776, 0.14576932787895203], "decide_seconds": 0.1735942690029333, "probabilities_f64_minus_runtime_max_abs": 2.6813884779208763e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.022195129083964198, "ibn_sina": 0.01728558391096223, "hypatia": 0.07746861248817094, "einstein": 0.832866074367365, "sagan": 0.04698708863947267, "darwin": 0.0031975115100649695}, "confidence": 0.6276378559033935, "choice": "einstein"}}, "usage": {"input_tokens": 316, "output_tokens": 1}, "latency_ms": 177.4369980012125, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 316, "compiled_prompt_sha256": "d6853c19ae77d785ff4d0e26d079aa1237f28072b8649be382018b78c39bc252", "compiled_ids_sha256": "c50647ddb58bcc8dc8ee0fc4a330c94e2ecf055f136663f079dabdd1c425c3eb", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 16.125, 17.625, 20.0, 17.125, 14.4375], "runtime_probabilities": [0.022195128723978996, 0.017285583540797234, 0.07746861129999161, 0.8328660726547241, 0.04698709025979042, 0.0031975116580724716], "decide_seconds": 0.17358233900085906, "probabilities_f64_minus_runtime_max_abs": 1.7126409179724078e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:06Z"} {"id": "c-sagan-01-main-03", "uid": "c083", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "prompt_sha256_by_order": ["8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad"], "request_sha256": "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "compiled_prompt_sha256_by_order": ["24b38735ab28fcd79f74277c7c0d9977f3ca86b28f20b968104f2e28ecd092e0", "ce28dea2a0a54fcc0e167a92ae8994c68835cad18971eebd17e64ab5f569e2ef", "77511413acf0a4b81abae4ee583725750c165d808590e86e0620d3a48e4a0cd4", "43f42d70a8389070f921aa5b565f8a3579f4d4a0689dd5a24d78d8e1582ee5b0", "f717a4e8245349623daaaf57dd81e85e8bbbbfb3c80e328ea9dd74481098fbe9", "fafbafb95e330679ee2c3ab99b2174edbe94237eb1806c2fe432e0d7decbe8ce"], "compiled_tokens_by_order": [399, 399, 399, 399, 399, 399], "choice": "sagan", "probs": {"darwin": 0.03654225054567907, "einstein": 0.0773599450122671, "hypatia": 0.028459133340166327, "ibn_sina": 0.017261336919660884, "sagan": 0.8316977889010254, "socrates": 0.008679545281201143}, "confidence": 0.617834740827609, "seconds": 0.21256758683315033, "seconds_all_orders": 1.275405520998902, "seconds_by_order": [0.21233058700090623, 0.21248555599959218, 0.2126053249994584, 0.21250716600115993, 0.21290177199989557, 0.2125751149978896], "argmax_stable": false, "spread": {"darwin": 0.03634092686145231, "einstein": 0.524044099640461, "hypatia": 0.03703350650580168, "ibn_sina": 0.02962677677714368, "sagan": 0.5436603734244242, "socrates": 0.024968215786140394}, "spread_2dp": {"darwin": 0.04, "einstein": 0.52, "hypatia": 0.03, "ibn_sina": 0.03, "sagan": 0.54, "socrates": 0.019999999999999997}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.03654225054567907, "einstein": 0.0773599450122671, "hypatia": 0.028459133340166327, "ibn_sina": 0.017261336919660884, "sagan": 0.8316977889010254, "socrates": 0.008679545281201143}, {"ibn_sina": 0.02027314812527888, "hypatia": 0.006581727559993091, "sagan": 0.8620361938344023, "socrates": 0.010851434226113862, "darwin": 0.029497256469563853, "einstein": 0.07076023978464788}, {"hypatia": 0.017880438351467527, "darwin": 0.013081608150350572, "sagan": 0.7602955853398182, "ibn_sina": 0.02601588429099528, "einstein": 0.16964487571701783, "socrates": 0.013081608150350572}, {"ibn_sina": 0.04361523406579477, "einstein": 0.1725015970183575, "darwin": 0.04942253501180288, "sagan": 0.6822570510339314, "hypatia": 0.04361523406579477, "socrates": 0.00858834880431868}, {"einstein": 0.09659330013672099, "sagan": 0.8087658000846741, "socrates": 0.019020348543423236, "ibn_sina": 0.021552878528771902, "darwin": 0.04562744420346108, "hypatia": 0.008440228502948646}, {"socrates": 0.033556564590459075, "ibn_sina": 0.013988457288651088, "hypatia": 0.026133878780235705, "einstein": 0.5948043394251088, "sagan": 0.31837582040997814, "darwin": 0.013140939505567207}], "prompt_tokens": 399, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.03654225054567907, "einstein": 0.0773599450122671, "hypatia": 0.028459133340166327, "ibn_sina": 0.017261336919660884, "sagan": 0.8316977889010254, "socrates": 0.008679545281201143}, "confidence": 0.617834740827609, "choice": "sagan"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 212.30297700094525, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 399, "compiled_prompt_sha256": "24b38735ab28fcd79f74277c7c0d9977f3ca86b28f20b968104f2e28ecd092e0", "compiled_ids_sha256": "c413e920def965f2042bf7df2f36f935761ed11a9ea1a07f8d6b709f90857e2d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 17.625, 16.625, 16.125, 20.0, 15.4375], "runtime_probabilities": [0.03654225170612335, 0.07735994458198547, 0.028459133580327034, 0.017261337488889694, 0.8316978216171265, 0.008679545484483242], "decide_seconds": 0.20754445399870747, "probabilities_f64_minus_runtime_max_abs": 3.27161010504895e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.02027314812527888, "hypatia": 0.006581727559993091, "sagan": 0.8620361938344023, "socrates": 0.010851434226113862, "darwin": 0.029497256469563853, "einstein": 0.07076023978464788}, "confidence": 0.6760183924946886, "choice": "sagan"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 212.46150599836255, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 399, "compiled_prompt_sha256": "ce28dea2a0a54fcc0e167a92ae8994c68835cad18971eebd17e64ab5f569e2ef", "compiled_ids_sha256": "bf0094bac882a568df3304d91bbb1ca238e146cfd92ef576cc50e57d22116f4e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 15.375, 20.25, 15.875, 16.875, 17.75], "runtime_probabilities": [0.020273147150874138, 0.006581726484000683, 0.8620361089706421, 0.010851433500647545, 0.02949725277721882, 0.07076023519039154], "decide_seconds": 0.2077829930021835, "probabilities_f64_minus_runtime_max_abs": 8.486376024130493e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.017880438351467527, "darwin": 0.013081608150350572, "sagan": 0.7602955853398182, "ibn_sina": 0.02601588429099528, "einstein": 0.16964487571701783, "socrates": 0.013081608150350572}, "confidence": 0.5592831292955749, "choice": "sagan"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 212.57991500169737, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 399, "compiled_prompt_sha256": "77511413acf0a4b81abae4ee583725750c165d808590e86e0620d3a48e4a0cd4", "compiled_ids_sha256": "1840e4a628ef8a1d24dc10b679fe046b93574447ba930937e0a19cfe31b59fa5", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 15.9375, 20.0, 16.625, 18.5, 15.9375], "runtime_probabilities": [0.017880437895655632, 0.013081607408821583, 0.760295569896698, 0.026015883311629295, 0.16964487731456757, 0.013081607408821583], "decide_seconds": 0.20792336099839304, "probabilities_f64_minus_runtime_max_abs": 1.5443120227054408e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.04361523406579477, "einstein": 0.1725015970183575, "darwin": 0.04942253501180288, "sagan": 0.6822570510339314, "hypatia": 0.04361523406579477, "socrates": 0.00858834880431868}, "confidence": 0.4269708262764308, "choice": "sagan"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 212.48185600052238, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 399, "compiled_prompt_sha256": "43f42d70a8389070f921aa5b565f8a3579f4d4a0689dd5a24d78d8e1582ee5b0", "compiled_ids_sha256": "0765e26d6f121f66398db5c61a9d90108bf60fa208ecde1dc2c3d3b33800173e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 18.5, 17.25, 19.875, 17.125, 15.5], "runtime_probabilities": [0.04361523315310478, 0.1725016087293625, 0.04942253977060318, 0.6822570562362671, 0.04361523315310478, 0.00858834944665432], "decide_seconds": 0.20764806299848715, "probabilities_f64_minus_runtime_max_abs": 1.1711004999082064e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.09659330013672099, "sagan": 0.8087658000846741, "socrates": 0.019020348543423236, "ibn_sina": 0.021552878528771902, "darwin": 0.04562744420346108, "hypatia": 0.008440228502948646}, "confidence": 0.5888680323961836, "choice": "sagan"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 212.87621200099238, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 399, "compiled_prompt_sha256": "f717a4e8245349623daaaf57dd81e85e8bbbbfb3c80e328ea9dd74481098fbe9", "compiled_ids_sha256": "aea9ce8aef233bdbabac501091443650f0703f298c762ae00bddccb5f19e513b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.125, 20.25, 16.5, 16.625, 17.375, 15.6875], "runtime_probabilities": [0.09659330546855927, 0.8087658286094666, 0.01902034878730774, 0.021552879363298416, 0.045627448707818985, 0.00844022911041975], "decide_seconds": 0.20817618000000948, "probabilities_f64_minus_runtime_max_abs": 2.8524792416995126e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.033556564590459075, "ibn_sina": 0.013988457288651088, "hypatia": 0.026133878780235705, "einstein": 0.5948043394251088, "sagan": 0.31837582040997814, "darwin": 0.013140939505567207}, "confidence": 0.44233156580500654, "choice": "einstein"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 212.55036500224378, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 399, "compiled_prompt_sha256": "fafbafb95e330679ee2c3ab99b2174edbe94237eb1806c2fe432e0d7decbe8ce", "compiled_ids_sha256": "2982ffbcaea14e2fe889c678fa398fb5e9810eba97ed7f717e43a428934f4889", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 16.0, 16.625, 19.75, 19.125, 15.9375], "runtime_probabilities": [0.03355656564235687, 0.01398845762014389, 0.026133878156542778, 0.5948043465614319, 0.318375825881958, 0.013140938244760036], "decide_seconds": 0.20790283099995577, "probabilities_f64_minus_runtime_max_abs": 7.1363230880194806e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:07Z"} {"id": "c-sagan-01-main-03", "uid": "c084", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "prompt_sha256_by_order": ["6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726"], "request_sha256": "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "compiled_prompt_sha256_by_order": ["18fc51ab32b6e239ab8298165343e7ff14bc85e0d166e6dc0b5e9964cae0940b", "55e53c7efb5da7ac634f55373a8eede9f8afa98982adeda8d85d7fdf5b2f8305", "38f455cfd986d076d286f2f4dac7050010666dd4fe22f1ff3e497ecfd95f5e8e", "a1ffcaa9ca0dbb7f9a7d2ff5069d542627fbde9107efb777269f8e5f71022423", "6ba87c856e7f4a01dce52756c360eb546ab93e9bd4f614569d416638894f9a4f", "380b1f584093c36086b4f2f4ddc88cc74acf880709274ca41d30bdb74ddc8d43"], "compiled_tokens_by_order": [400, 400, 400, 400, 400, 400], "choice": "sagan", "probs": {"darwin": 0.014872277935245867, "einstein": 0.15989205618882077, "hypatia": 0.0058240677209434885, "ibn_sina": 0.0048283217269167135, "sagan": 0.8119988620431745, "socrates": 0.002584414384898718}, "confidence": 0.6674040687579341, "seconds": 0.2122795793335778, "seconds_all_orders": 1.2736774760014669, "seconds_by_order": [0.21227299800011679, 0.21206249900205876, 0.21226149800349958, 0.21265535499696853, 0.2125257359984971, 0.2118993900003261], "argmax_stable": true, "spread": {"darwin": 0.018702333834669196, "einstein": 0.1195033520201185, "hypatia": 0.022643783364538872, "ibn_sina": 0.014233362367245105, "sagan": 0.10871923741600198, "socrates": 0.020748987027712443}, "spread_2dp": {"darwin": 0.01, "einstein": 0.12, "hypatia": 0.03, "ibn_sina": 0.02, "sagan": 0.10999999999999999, "socrates": 0.02}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.014872277935245867, "einstein": 0.15989205618882077, "hypatia": 0.0058240677209434885, "ibn_sina": 0.0048283217269167135, "sagan": 0.8119988620431745, "socrates": 0.002584414384898718}, {"ibn_sina": 0.00846593230603701, "hypatia": 0.0025819634744155702, "sagan": 0.9192426698431492, "socrates": 0.004823742823314545, "darwin": 0.024496987384381372, "einstein": 0.04038870416870227}, {"hypatia": 0.005443573239078093, "darwin": 0.005794653549712175, "sagan": 0.9154682570009145, "ibn_sina": 0.008974934987882047, "einstein": 0.05852392767270096, "socrates": 0.005794653549712175}, {"ibn_sina": 0.01906168409416182, "einstein": 0.12429779410690278, "darwin": 0.024475686861776223, "sagan": 0.8105234324271472, "hypatia": 0.01906168409416182, "socrates": 0.002579718415850225}, {"einstein": 0.05596265512879823, "sagan": 0.8754032134415659, "socrates": 0.023328705443562667, "ibn_sina": 0.01816841406748875, "darwin": 0.023328705443562667, "hypatia": 0.0038083064750218366}, {"socrates": 0.013502369288477157, "ibn_sina": 0.0056286249320396525, "hypatia": 0.02522574683895444, "einstein": 0.11305395389925713, "sagan": 0.8353620075675303, "darwin": 0.007227297473741225}], "prompt_tokens": 400, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.014872277935245867, "einstein": 0.15989205618882077, "hypatia": 0.0058240677209434885, "ibn_sina": 0.0048283217269167135, "sagan": 0.8119988620431745, "socrates": 0.002584414384898718}, "confidence": 0.6674040687579341, "choice": "sagan"}}, "usage": {"input_tokens": 400, "output_tokens": 1}, "latency_ms": 212.2423879991402, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 400, "compiled_prompt_sha256": "18fc51ab32b6e239ab8298165343e7ff14bc85e0d166e6dc0b5e9964cae0940b", "compiled_ids_sha256": "73adbcdaa744c056b6a9cf97351cb9d64087fb324180be15d21ff25d3141cbfd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 18.375, 15.0625, 14.875, 20.0, 14.25], "runtime_probabilities": [0.014872279018163681, 0.15989205241203308, 0.005824067629873753, 0.004828321747481823, 0.8119989037513733, 0.002584414556622505], "decide_seconds": 0.20741198500036262, "probabilities_f64_minus_runtime_max_abs": 4.1708198783574346e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.00846593230603701, "hypatia": 0.0025819634744155702, "sagan": 0.9192426698431492, "socrates": 0.004823742823314545, "darwin": 0.024496987384381372, "einstein": 0.04038870416870227}, "confidence": 0.7882532995496809, "choice": "sagan"}}, "usage": {"input_tokens": 400, "output_tokens": 1}, "latency_ms": 212.03544899981353, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 400, "compiled_prompt_sha256": "55e53c7efb5da7ac634f55373a8eede9f8afa98982adeda8d85d7fdf5b2f8305", "compiled_ids_sha256": "082cb715241a64ca4a2624f19ffc2130c7fb6a9bfe2ee77e84249a9596f63472", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5625, 14.375, 20.25, 15.0, 16.625, 17.125], "runtime_probabilities": [0.008465931750833988, 0.0025819637812674046, 0.9192426800727844, 0.004823742900043726, 0.024496987462043762, 0.04038870707154274], "decide_seconds": 0.2074323250017187, "probabilities_f64_minus_runtime_max_abs": 1.0229635249991986e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.005443573239078093, "darwin": 0.005794653549712175, "sagan": 0.9154682570009145, "ibn_sina": 0.008974934987882047, "einstein": 0.05852392767270096, "socrates": 0.005794653549712175}, "confidence": 0.7894032930956882, "choice": "sagan"}}, "usage": {"input_tokens": 400, "output_tokens": 1}, "latency_ms": 212.2340779969818, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 400, "compiled_prompt_sha256": "38f455cfd986d076d286f2f4dac7050010666dd4fe22f1ff3e497ecfd95f5e8e", "compiled_ids_sha256": "85e9a30fa7702e0d8709a5785cef96e67c06280ab7efd776837cf4537b2873c3", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.25, 15.3125, 20.375, 15.75, 17.625, 15.3125], "runtime_probabilities": [0.005443572998046875, 0.005794653668999672, 0.9154682755470276, 0.008974935859441757, 0.05852392688393593, 0.005794653668999672], "decide_seconds": 0.20741723500032094, "probabilities_f64_minus_runtime_max_abs": 1.8546113089712435e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.01906168409416182, "einstein": 0.12429779410690278, "darwin": 0.024475686861776223, "sagan": 0.8105234324271472, "hypatia": 0.01906168409416182, "socrates": 0.002579718415850225}, "confidence": 0.6168044212942931, "choice": "sagan"}}, "usage": {"input_tokens": 400, "output_tokens": 1}, "latency_ms": 212.6282950011955, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 400, "compiled_prompt_sha256": "a1ffcaa9ca0dbb7f9a7d2ff5069d542627fbde9107efb777269f8e5f71022423", "compiled_ids_sha256": "74d152d484047691c38ac5fdebfb1e0524429e0edb578d3126afff2e3aefc7c7", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 18.0, 16.375, 19.875, 16.125, 14.125], "runtime_probabilities": [0.019061684608459473, 0.12429779767990112, 0.024475688114762306, 0.8105234503746033, 0.019061684608459473, 0.0025797185953706503], "decide_seconds": 0.20792412100126967, "probabilities_f64_minus_runtime_max_abs": 1.79474560768611e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.05596265512879823, "sagan": 0.8754032134415659, "socrates": 0.023328705443562667, "ibn_sina": 0.01816841406748875, "darwin": 0.023328705443562667, "hypatia": 0.0038083064750218366}, "confidence": 0.6945955191849746, "choice": "sagan"}}, "usage": {"input_tokens": 400, "output_tokens": 1}, "latency_ms": 212.49854599955142, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 400, "compiled_prompt_sha256": "6ba87c856e7f4a01dce52756c360eb546ab93e9bd4f614569d416638894f9a4f", "compiled_ids_sha256": "7cc99a50550515f05096a6fe74c9707656dbd6595edb8c239ebc990f28b21f5f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 20.125, 16.5, 16.25, 16.5, 14.6875], "runtime_probabilities": [0.055962659418582916, 0.8754032850265503, 0.023328706622123718, 0.018168415874242783, 0.023328706622123718, 0.0038083067629486322], "decide_seconds": 0.20789967100063222, "probabilities_f64_minus_runtime_max_abs": 7.158498438464278e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.013502369288477157, "ibn_sina": 0.0056286249320396525, "hypatia": 0.02522574683895444, "einstein": 0.11305395389925713, "sagan": 0.8353620075675303, "darwin": 0.007227297473741225}, "confidence": 0.6581805765142266, "choice": "sagan"}}, "usage": {"input_tokens": 400, "output_tokens": 1}, "latency_ms": 211.8725110012747, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 400, "compiled_prompt_sha256": "380b1f584093c36086b4f2f4ddc88cc74acf880709274ca41d30bdb74ddc8d43", "compiled_ids_sha256": "f3ec31b87dd977f9e983b20c34c4b2e7dced1ed7adf3d950f5d6ee2845f9b206", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.875, 15.0, 16.5, 18.0, 20.0, 15.25], "runtime_probabilities": [0.01350236963480711, 0.005628625396639109, 0.025225749239325523, 0.11305395513772964, 0.8353620767593384, 0.007227297872304916], "decide_seconds": 0.2072187660014606, "probabilities_f64_minus_runtime_max_abs": 6.919180806086445e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:08Z"} {"id": "c-sagan-01-starter-01", "uid": "c085", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "prompt_sha256_by_order": ["895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655"], "request_sha256": "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "compiled_prompt_sha256_by_order": ["65b093d4dc7322a71d795f378d96f989360cca3b9cc8cc2dafc040037e482792", "29a9bb31b220de972a1c5ad69356b5d1101174b43840ce57d6315d8082c32b05", "ce7c4dc2c035eee38fbc626a23e11fb9ededff2b103befafff23d764bfad4c17", "f37681dc34ec37cec21c16f17078a46c8da2e6e2793ff9b7f85f423475ce1c3e", "152f3406252b1770dd3a9f9aa80f49ee5566c556d5802738b80d336bff5724bf", "38f69caf3a1bd95791d66f62f117a483a70022cbbb21d3824557a66bd1a6fc15"], "compiled_tokens_by_order": [284, 284, 284, 284, 284, 284], "choice": "sagan", "probs": {"darwin": 0.004273698393701086, "einstein": 0.010913271165095557, "hypatia": 0.001080560174454777, "ibn_sina": 0.0009535910070126107, "sagan": 0.9823813633866054, "socrates": 0.0003975158731305396}, "confidence": 0.940167139662661, "seconds": 0.17360667999976917, "seconds_all_orders": 1.041640079998615, "seconds_by_order": [0.17381082700012485, 0.17388440600188915, 0.1737223569980415, 0.17357380899920827, 0.17351355899882037, 0.17313512200053083], "argmax_stable": true, "spread": {"darwin": 0.004656231971122688, "einstein": 0.009173821361487033, "hypatia": 0.00290573118929305, "ibn_sina": 0.001622840324393578, "sagan": 0.014531044632890677, "socrates": 0.003729700606430175}, "spread_2dp": {"darwin": 0.01, "einstein": 0.01, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.010000000000000009, "socrates": 0.0}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.004273698393701086, "einstein": 0.010913271165095557, "hypatia": 0.001080560174454777, "ibn_sina": 0.0009535910070126107, "sagan": 0.9823813633866054, "socrates": 0.0003975158731305396}, {"ibn_sina": 0.0025764313314061886, "hypatia": 0.0010089450061250601, "sagan": 0.9764321221575137, "socrates": 0.0018849574289645725, "darwin": 0.005806077632189889, "einstein": 0.012291466443800473}, {"hypatia": 0.0014764310540784046, "darwin": 0.0011498456610672012, "sagan": 0.9820357623871185, "ibn_sina": 0.001895774999423745, "einstein": 0.012362005864086164, "socrates": 0.0010801800342259253}, {"ibn_sina": 0.0024263148126278805, "einstein": 0.015821560022124058, "darwin": 0.001382472507625594, "sagan": 0.9788452574920176, "hypatia": 0.0012987127326439855, "socrates": 0.000225682432960849}, {"einstein": 0.01236519464938243, "sagan": 0.9822890789794205, "socrates": 0.0013873363903395006, "ibn_sina": 0.001150142264302736, "darwin": 0.0024348511926358314, "hypatia": 0.00037339652391900243}, {"socrates": 0.003955383039391024, "ibn_sina": 0.0015489502538875042, "hypatia": 0.0032791277132120525, "einstein": 0.02008709252658259, "sagan": 0.9678503187537147, "darwin": 0.0032791277132120525}], "prompt_tokens": 284, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.004273698393701086, "einstein": 0.010913271165095557, "hypatia": 0.001080560174454777, "ibn_sina": 0.0009535910070126107, "sagan": 0.9823813633866054, "socrates": 0.0003975158731305396}, "confidence": 0.940167139662661, "choice": "sagan"}}, "usage": {"input_tokens": 284, "output_tokens": 1}, "latency_ms": 173.784586997499, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 284, "compiled_prompt_sha256": "65b093d4dc7322a71d795f378d96f989360cca3b9cc8cc2dafc040037e482792", "compiled_ids_sha256": "c711190daf3352b61206e8801960f56744ba55154d91e3a74652ef38d6ba1604", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.9375, 15.875, 13.5625, 13.4375, 20.375, 12.5625], "runtime_probabilities": [0.004273698665201664, 0.010913272388279438, 0.0010805601486936212, 0.0009535910212434828, 0.9823814034461975, 0.0003975158906541765], "decide_seconds": 0.1701966850014287, "probabilities_f64_minus_runtime_max_abs": 4.005959208797094e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0025764313314061886, "hypatia": 0.0010089450061250601, "sagan": 0.9764321221575137, "socrates": 0.0018849574289645725, "darwin": 0.005806077632189889, "einstein": 0.012291466443800473}, "confidence": 0.9210851677894335, "choice": "sagan"}}, "usage": {"input_tokens": 284, "output_tokens": 1}, "latency_ms": 173.8616459988407, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 284, "compiled_prompt_sha256": "29a9bb31b220de972a1c5ad69356b5d1101174b43840ce57d6315d8082c32b05", "compiled_ids_sha256": "b34c05271926877c4d55edccdd464343b298f3cb911940c7a6a56c5f09eb4aee", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.8125, 13.875, 20.75, 14.5, 15.625, 16.375], "runtime_probabilities": [0.0025764312595129013, 0.0010089449351653457, 0.9764320850372314, 0.0018849573098123074, 0.005806077271699905, 0.012291465885937214], "decide_seconds": 0.17015624600026058, "probabilities_f64_minus_runtime_max_abs": 3.7120282292413265e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0014764310540784046, "darwin": 0.0011498456610672012, "sagan": 0.9820357623871185, "ibn_sina": 0.001895774999423745, "einstein": 0.012362005864086164, "socrates": 0.0010801800342259253}, "confidence": 0.9392904509795531, "choice": "sagan"}}, "usage": {"input_tokens": 284, "output_tokens": 1}, "latency_ms": 173.7005870018038, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 284, "compiled_prompt_sha256": "ce7c4dc2c035eee38fbc626a23e11fb9ededff2b103befafff23d764bfad4c17", "compiled_ids_sha256": "429fe0a0e84aa69822311049d513022e69e24731fa215f45eff10b1d3c6ce3b3", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.375, 14.125, 20.875, 14.625, 16.5, 14.0625], "runtime_probabilities": [0.0014764309162274003, 0.0011498455423861742, 0.9820356965065002, 0.0018957749707624316, 0.012362005189061165, 0.0010801799362525344], "decide_seconds": 0.17018421499960823, "probabilities_f64_minus_runtime_max_abs": 6.588061829404523e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0024263148126278805, "einstein": 0.015821560022124058, "darwin": 0.001382472507625594, "sagan": 0.9788452574920176, "hypatia": 0.0012987127326439855, "socrates": 0.000225682432960849}, "confidence": 0.9325969390128108, "choice": "sagan"}}, "usage": {"input_tokens": 284, "output_tokens": 1}, "latency_ms": 173.55267899984028, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 284, "compiled_prompt_sha256": "f37681dc34ec37cec21c16f17078a46c8da2e6e2793ff9b7f85f423475ce1c3e", "compiled_ids_sha256": "b7269e259ce88d0455d1f3329a83764530595bc0df657ec82d7fbcf0ead6697e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.625, 16.5, 14.0625, 20.625, 14.0, 12.25], "runtime_probabilities": [0.002426314866170287, 0.015821559354662895, 0.0013824725756421685, 0.9788452386856079, 0.0012987126829102635, 0.0002256824227515608], "decide_seconds": 0.17016718500235584, "probabilities_f64_minus_runtime_max_abs": 1.880640965179481e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.01236519464938243, "sagan": 0.9822890789794205, "socrates": 0.0013873363903395006, "ibn_sina": 0.001150142264302736, "darwin": 0.0024348511926358314, "hypatia": 0.00037339652391900243}, "confidence": 0.940625499842865, "choice": "sagan"}}, "usage": {"input_tokens": 284, "output_tokens": 1}, "latency_ms": 173.49183999976958, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 284, "compiled_prompt_sha256": "152f3406252b1770dd3a9f9aa80f49ee5566c556d5802738b80d336bff5724bf", "compiled_ids_sha256": "a29acbd8841f67c53988645b61dd30d45fa4aa082bdad89c04b9d96ca097a6ac", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 20.75, 14.1875, 14.0, 14.75, 12.875], "runtime_probabilities": [0.012365194037556648, 0.9822890758514404, 0.0013873364077880979, 0.0011501421686261892, 0.002434851136058569, 0.00037339652772061527], "decide_seconds": 0.17006859600223834, "probabilities_f64_minus_runtime_max_abs": 3.1279800971617533e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.003955383039391024, "ibn_sina": 0.0015489502538875042, "hypatia": 0.0032791277132120525, "einstein": 0.02008709252658259, "sagan": 0.9678503187537147, "darwin": 0.0032791277132120525}, "confidence": 0.8997960544445888, "choice": "sagan"}}, "usage": {"input_tokens": 284, "output_tokens": 1}, "latency_ms": 173.1145119993016, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 284, "compiled_prompt_sha256": "38f69caf3a1bd95791d66f62f117a483a70022cbbb21d3824557a66bd1a6fc15", "compiled_ids_sha256": "26d971bdc909a84a904f4e9c530caa375291f3369e7a3de38e9e1d25f10ee900", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.75, 13.8125, 14.5625, 16.375, 20.25, 14.5625], "runtime_probabilities": [0.0039553833194077015, 0.0015489504439756274, 0.0032791278790682554, 0.020087093114852905, 0.967850387096405, 0.0032791278790682554], "decide_seconds": 0.16972183899997617, "probabilities_f64_minus_runtime_max_abs": 6.83426902847728e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:09Z"} {"id": "c-sagan-01-starter-02", "uid": "c086", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "prompt_sha256_by_order": ["30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263"], "request_sha256": "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "compiled_prompt_sha256_by_order": ["e2b81005504d2eebe6b4f6eae69c57ba10486928d37eddc53c160a4b2854c594", "9f7e16f19df496ad1969d45d35fdba9da41a87f38a37323bc74379e8bafeaa71", "6403a7bcb3e90824947545b04e0ce62f413fe159bd666c42fbb3771e5d9f598f", "b6b13f14b5cd2091588d3644643557e4d50a01260a02fce23b08a21c3144f3c8", "b1eddab09bd632438871f1817cb5d6b314a85f8d31b1c6c3bfa99dbc7399640c", "230ef714e3cb334d5bc4c4ea4dff194999629797df91836c38b3f68ebc471b30"], "compiled_tokens_by_order": [404, 404, 404, 404, 404, 404], "choice": "einstein", "probs": {"darwin": 0.043584438352172475, "einstein": 0.36492803440034255, "hypatia": 0.06341498361336723, "ibn_sina": 0.36492803440034255, "sagan": 0.15212463745743499, "socrates": 0.01101987177634018}, "confidence": 0.2279505723633397, "seconds": 0.2134533251658771, "seconds_all_orders": 1.2807199509952625, "seconds_by_order": [0.2137701460014796, 0.21382017599898973, 0.21332538899878273, 0.213282459997572, 0.21280974299952504, 0.2137120369989134], "argmax_stable": false, "spread": {"darwin": 0.18235591926781267, "einstein": 0.5420955967293479, "hypatia": 0.039860539562389344, "ibn_sina": 0.37325026893091906, "sagan": 0.3209790182336393, "socrates": 0.0545016923823147}, "spread_2dp": {"darwin": 0.19, "einstein": 0.5499999999999999, "hypatia": 0.039999999999999994, "ibn_sina": 0.37, "sagan": 0.31999999999999995, "socrates": 0.049999999999999996}, "choices_by_order": ["einstein", "sagan", "einstein", "einstein", "ibn_sina", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.043584438352172475, "einstein": 0.36492803440034255, "hypatia": 0.06341498361336723, "ibn_sina": 0.36492803440034255, "sagan": 0.15212463745743499, "socrates": 0.01101987177634018}, {"ibn_sina": 0.08221308791025961, "hypatia": 0.02355444405097788, "sagan": 0.47310365569107427, "socrates": 0.04986475844722248, "darwin": 0.19721894542866686, "einstein": 0.17404510847179888}, {"hypatia": 0.037360111412070164, "darwin": 0.03297018260136732, "sagan": 0.24361852883349036, "ibn_sina": 0.27605595908608566, "einstein": 0.40165905042444805, "socrates": 0.008336167642538396}, {"ibn_sina": 0.06235839953266485, "einstein": 0.40662784398578533, "darwin": 0.05503109443770942, "sagan": 0.40662784398578533, "hypatia": 0.06235839953266485, "socrates": 0.006996418525390253}, {"einstein": 0.21464949830391522, "sagan": 0.16716919736497335, "socrates": 0.06149811090770495, "ibn_sina": 0.4010180574711443, "darwin": 0.1013932435614087, "hypatia": 0.05427189239085354}, {"socrates": 0.035654546250859794, "ibn_sina": 0.02776778854022523, "hypatia": 0.04578134360657259, "einstein": 0.7161407052011468, "sagan": 0.15979259024034137, "darwin": 0.014863026160854203}], "prompt_tokens": 404, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.043584438352172475, "einstein": 0.36492803440034255, "hypatia": 0.06341498361336723, "ibn_sina": 0.36492803440034255, "sagan": 0.15212463745743499, "socrates": 0.01101987177634018}, "confidence": 0.2279505723633397, "choice": "einstein"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 213.74157600075705, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 404, "compiled_prompt_sha256": "e2b81005504d2eebe6b4f6eae69c57ba10486928d37eddc53c160a4b2854c594", "compiled_ids_sha256": "6f3ebc0833ff0d4eca3500ecdd2bf006ccd0dd84cbb8ec6be2d67191859098b9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 19.0, 17.25, 19.0, 18.125, 15.5], "runtime_probabilities": [0.043584439903497696, 0.3649280369281769, 0.06341497600078583, 0.3649280369281769, 0.15212462842464447, 0.011019871570169926], "decide_seconds": 0.2088946939984453, "probabilities_f64_minus_runtime_max_abs": 9.032790515561828e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.08221308791025961, "hypatia": 0.02355444405097788, "sagan": 0.47310365569107427, "socrates": 0.04986475844722248, "darwin": 0.19721894542866686, "einstein": 0.17404510847179888}, "confidence": 0.20648701909117184, "choice": "sagan"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 213.79439600059413, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 404, "compiled_prompt_sha256": "9f7e16f19df496ad1969d45d35fdba9da41a87f38a37323bc74379e8bafeaa71", "compiled_ids_sha256": "ebc29f3be1769f4245904f8ffdbe8182607653d61981ed695aa218af9787ded6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 16.375, 19.375, 17.125, 18.5, 18.375], "runtime_probabilities": [0.08221307396888733, 0.023554442450404167, 0.4731036424636841, 0.0498647540807724, 0.19721892476081848, 0.1740451157093048], "decide_seconds": 0.20879146399965975, "probabilities_f64_minus_runtime_max_abs": 2.0667848377176057e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.037360111412070164, "darwin": 0.03297018260136732, "sagan": 0.24361852883349036, "ibn_sina": 0.27605595908608566, "einstein": 0.40165905042444805, "socrates": 0.008336167642538396}, "confidence": 0.25160696715736297, "choice": "einstein"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 213.30062999913935, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 404, "compiled_prompt_sha256": "6403a7bcb3e90824947545b04e0ce62f413fe159bd666c42fbb3771e5d9f598f", "compiled_ids_sha256": "6e2c14c80d83f33fd3c4523a18516f272ecdf355dada430a4e086c4467b2d1a5", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 16.75, 18.75, 18.875, 19.25, 15.375], "runtime_probabilities": [0.03736010938882828, 0.032970182597637177, 0.24361851811408997, 0.2760559618473053, 0.4016590416431427, 0.008336166851222515], "decide_seconds": 0.20856557599836378, "probabilities_f64_minus_runtime_max_abs": 1.0719400395098333e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.06235839953266485, "einstein": 0.40662784398578533, "darwin": 0.05503109443770942, "sagan": 0.40662784398578533, "hypatia": 0.06235839953266485, "socrates": 0.006996418525390253}, "confidence": 0.2899794535372501, "choice": "einstein"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 213.25781999985338, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 404, "compiled_prompt_sha256": "b6b13f14b5cd2091588d3644643557e4d50a01260a02fce23b08a21c3144f3c8", "compiled_ids_sha256": "e9fea21702e5747b46432936efcbde9d65200b9e1bc6e57d7471bd0a9431b79a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 19.0, 17.0, 19.0, 17.125, 14.9375], "runtime_probabilities": [0.06235840171575546, 0.4066278338432312, 0.05503109470009804, 0.4066278338432312, 0.06235840171575546, 0.006996418349444866], "decide_seconds": 0.20854944600068848, "probabilities_f64_minus_runtime_max_abs": 1.0142554129899395e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.21464949830391522, "sagan": 0.16716919736497335, "socrates": 0.06149811090770495, "ibn_sina": 0.4010180574711443, "darwin": 0.1013932435614087, "hypatia": 0.05427189239085354}, "confidence": 0.13077269293003246, "choice": "ibn_sina"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 212.78458400047384, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 404, "compiled_prompt_sha256": "b1eddab09bd632438871f1817cb5d6b314a85f8d31b1c6c3bfa99dbc7399640c", "compiled_ids_sha256": "f49f16cb89f90e7089d843c9b1c735f8b93a6df46048d0a8d32b71d189ed425b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.375, 18.125, 17.125, 19.0, 17.625, 17.0], "runtime_probabilities": [0.2146495133638382, 0.16716919839382172, 0.061498112976551056, 0.40101805329322815, 0.10139324516057968, 0.054271891713142395], "decide_seconds": 0.20813620900298702, "probabilities_f64_minus_runtime_max_abs": 1.505992297534675e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.035654546250859794, "ibn_sina": 0.02776778854022523, "hypatia": 0.04578134360657259, "einstein": 0.7161407052011468, "sagan": 0.15979259024034137, "darwin": 0.014863026160854203}, "confidence": 0.46741212186508596, "choice": "einstein"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 213.6871569964569, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 404, "compiled_prompt_sha256": "230ef714e3cb334d5bc4c4ea4dff194999629797df91836c38b3f68ebc471b30", "compiled_ids_sha256": "843cb8d51c5a7617174e7c789eddfe37c09f43daae11d33fb7de9a493746acf2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 16.5, 17.0, 19.75, 18.25, 15.875], "runtime_probabilities": [0.03565454110503197, 0.027767786756157875, 0.04578133672475815, 0.716140627861023, 0.15979258716106415, 0.014863024465739727], "decide_seconds": 0.2088682939975115, "probabilities_f64_minus_runtime_max_abs": 7.73401238607363e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:11Z"} {"id": "c-sagan-01-starter-02", "uid": "c087", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "prompt_sha256_by_order": ["dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a"], "request_sha256": "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "compiled_prompt_sha256_by_order": ["260bb13ff919a69d62276bf87d96d93384b67dfa42212c818966dc0bd3ee540c", "dc8f222ffc40c32ac80db98d97b1f8417747db7cd32c674f41221704a9d6ba61", "a0c465baebf66bf5c8365e0f484b66df78bb7fd324ee6383fce1fed93160f6ba", "a941b2dd0565dce93f296a6e61caa6b3a5a993226d375855a21cf35282d92283", "6b7d821c7e72bc68abb3b4f45a0262758e0eb531ab19382b8c49de1d06fb8603", "78775807212758e40edf769f9eb7a399a8186be9fa4466e97223edc44ad77b5e"], "compiled_tokens_by_order": [293, 293, 293, 293, 293, 293], "choice": "sagan", "probs": {"darwin": 0.01752350591932739, "einstein": 0.012043717745183855, "hypatia": 0.0077760086148914095, "ibn_sina": 0.0030451287266787383, "sagan": 0.956751005290128, "socrates": 0.002860633703790573}, "confidence": 0.8668591616827879, "seconds": 0.1746816964999501, "seconds_all_orders": 1.0480901789997006, "seconds_by_order": [0.17493569799989928, 0.17487610800162656, 0.17466155999863986, 0.1743084629997611, 0.17428538300009677, 0.175022966999677], "argmax_stable": true, "spread": {"darwin": 0.013935247308770618, "einstein": 0.037048047156087796, "hypatia": 0.0065358235295660035, "ibn_sina": 0.0022783923188845786, "sagan": 0.03287132891019828, "socrates": 0.006409635050787275}, "spread_2dp": {"darwin": 0.02, "einstein": 0.04, "hypatia": 0.01, "ibn_sina": 0.0, "sagan": 0.039999999999999925, "socrates": 0.01}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.01752350591932739, "einstein": 0.012043717745183855, "hypatia": 0.0077760086148914095, "ibn_sina": 0.0030451287266787383, "sagan": 0.956751005290128, "socrates": 0.002860633703790573}, {"ibn_sina": 0.002466402548403746, "hypatia": 0.0012401850853254056, "sagan": 0.9347327231570962, "socrates": 0.0033711725813659805, "darwin": 0.01712022701482851, "einstein": 0.0410692896129802}, {"hypatia": 0.004607415257603534, "darwin": 0.0035882586105567714, "sagan": 0.9346452243821917, "ibn_sina": 0.004328266078798875, "einstein": 0.04653324568601371, "socrates": 0.006297589984835551}, {"ibn_sina": 0.0034894095345893326, "einstein": 0.009485198529925916, "darwin": 0.010096941276841964, "sagan": 0.96751655329239, "hypatia": 0.0065190752568219735, "socrates": 0.0028928221094308756}, {"einstein": 0.022238218683791937, "sagan": 0.9455931201888452, "socrates": 0.009270268754577848, "ibn_sina": 0.003864422955995892, "darwin": 0.017319142125050306, "hypatia": 0.0017148272917387527}, {"socrates": 0.0071547624416026905, "ibn_sina": 0.0020498737599142967, "hypatia": 0.004917391518409849, "einstein": 0.04117281509263214, "sagan": 0.9370889522134814, "darwin": 0.007616204973959679}], "prompt_tokens": 293, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.01752350591932739, "einstein": 0.012043717745183855, "hypatia": 0.0077760086148914095, "ibn_sina": 0.0030451287266787383, "sagan": 0.956751005290128, "socrates": 0.002860633703790573}, "confidence": 0.8668591616827879, "choice": "sagan"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 174.908177999896, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 293, "compiled_prompt_sha256": "260bb13ff919a69d62276bf87d96d93384b67dfa42212c818966dc0bd3ee540c", "compiled_ids_sha256": "56677b4e63c956b87e199aa8482bbf7c197f78ec3d9d566703cd84fe0d74f83f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 15.875, 15.4375, 14.5, 20.25, 14.4375], "runtime_probabilities": [0.01752350479364395, 0.012043716385960579, 0.007776007987558842, 0.0030451284255832434, 0.9567508697509766, 0.002860633423551917], "decide_seconds": 0.17115772800025297, "probabilities_f64_minus_runtime_max_abs": 1.3553915145791962e-07}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.002466402548403746, "hypatia": 0.0012401850853254056, "sagan": 0.9347327231570962, "socrates": 0.0033711725813659805, "darwin": 0.01712022701482851, "einstein": 0.0410692896129802}, "confidence": 0.829139776519281, "choice": "sagan"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 174.85212799874716, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 293, "compiled_prompt_sha256": "dc8f222ffc40c32ac80db98d97b1f8417747db7cd32c674f41221704a9d6ba61", "compiled_ids_sha256": "af59a0ac3a4a9fd765bbceaec5d8a816f181f880d3da6a5d1d767a3b5090c0ed", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.3125, 13.625, 20.25, 14.625, 16.25, 17.125], "runtime_probabilities": [0.002466402482241392, 0.0012401849962770939, 0.9347326755523682, 0.003371172584593296, 0.017120227217674255, 0.041069287806749344], "decide_seconds": 0.17123720699964906, "probabilities_f64_minus_runtime_max_abs": 4.760472804843374e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.004607415257603534, "darwin": 0.0035882586105567714, "sagan": 0.9346452243821917, "ibn_sina": 0.004328266078798875, "einstein": 0.04653324568601371, "socrates": 0.006297589984835551}, "confidence": 0.8290077628404353, "choice": "sagan"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 174.63784999927157, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 293, "compiled_prompt_sha256": "a0c465baebf66bf5c8365e0f484b66df78bb7fd324ee6383fce1fed93160f6ba", "compiled_ids_sha256": "4d05d2b9672ba78fc309e45b366de89998c5443bfdcf55ee59864853fec964b9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.8125, 14.5625, 20.125, 14.75, 17.125, 15.125], "runtime_probabilities": [0.004607415292412043, 0.0035882587544620037, 0.9346451759338379, 0.004328266251832247, 0.046533241868019104, 0.006297589745372534], "decide_seconds": 0.1710921980011335, "probabilities_f64_minus_runtime_max_abs": 4.844835377326717e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0034894095345893326, "einstein": 0.009485198529925916, "darwin": 0.010096941276841964, "sagan": 0.96751655329239, "hypatia": 0.0065190752568219735, "socrates": 0.0028928221094308756}, "confidence": 0.8928444808694759, "choice": "sagan"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 174.28513300183113, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 293, "compiled_prompt_sha256": "a941b2dd0565dce93f296a6e61caa6b3a5a993226d375855a21cf35282d92283", "compiled_ids_sha256": "439ba1852928ce287bc969fdbf4ef486b2bd9cdc334691a62b1240ab5ee705dc", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.875, 15.875, 15.9375, 20.5, 15.5, 14.6875], "runtime_probabilities": [0.003489409340545535, 0.009485198184847832, 0.010096941143274307, 0.9675164818763733, 0.006519075017422438, 0.002892822027206421], "decide_seconds": 0.17073634099870105, "probabilities_f64_minus_runtime_max_abs": 7.141601665594521e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.022238218683791937, "sagan": 0.9455931201888452, "socrates": 0.009270268754577848, "ibn_sina": 0.003864422955995892, "darwin": 0.017319142125050306, "hypatia": 0.0017148272917387527}, "confidence": 0.8417382645816992, "choice": "sagan"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 174.26129300292814, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 293, "compiled_prompt_sha256": "6b7d821c7e72bc68abb3b4f45a0262758e0eb531ab19382b8c49de1d06fb8603", "compiled_ids_sha256": "b61a5716f9e55c7e3dd16078e1b806b15a10d5c0ddc7af3450078fb023c9391b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 20.375, 15.75, 14.875, 16.375, 14.0625], "runtime_probabilities": [0.022238219156861305, 0.9455931186676025, 0.00927026942372322, 0.003864422906190157, 0.017319142818450928, 0.0017148272600024939], "decide_seconds": 0.17067153099924326, "probabilities_f64_minus_runtime_max_abs": 1.5212426873745244e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.0071547624416026905, "ibn_sina": 0.0020498737599142967, "hypatia": 0.004917391518409849, "einstein": 0.04117281509263214, "sagan": 0.9370889522134814, "darwin": 0.007616204973959679}, "confidence": 0.8305875558486856, "choice": "sagan"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 174.99931700149318, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 293, "compiled_prompt_sha256": "78775807212758e40edf769f9eb7a399a8186be9fa4466e97223edc44ad77b5e", "compiled_ids_sha256": "7cec2fc6c5f8ca6a7e31daf6344b14e7480672c983adbefea1ea42e028662f59", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.125, 13.875, 14.75, 16.875, 20.0, 15.1875], "runtime_probabilities": [0.007154762279242277, 0.0020498738158494234, 0.004917391575872898, 0.0411728173494339, 0.9370890259742737, 0.00761620607227087], "decide_seconds": 0.17137348499818472, "probabilities_f64_minus_runtime_max_abs": 7.37607922518535e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:12Z"} {"id": "c-sagan-01-starter-03", "uid": "c088", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "prompt_sha256_by_order": ["ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f"], "request_sha256": "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "compiled_prompt_sha256_by_order": ["3824b179b4488ce489b4ef53aaa0bc5aac8823a7a8d88bcb5a7f0cf96ad47e1b", "dbc05fef85bcc0d56f0a06133eb4389f636ed6460a3f4b3cf3d4405cbf252d2a", "e452353fa315daca4512110c64c8bab66fbc44387cf2c7967e7b40996297fed1", "c668d5cb4b9609114af6406ec50b06bc347bf14cd986105eb07aab8d3ce32c19", "224425416d064d538ee268e1d83ab285dc072fbc957d6598f7ce41c93e35e9cf", "d761f71fff8dabb20c919c88489e7ab111f0c5dc12bd69b7aa2258f047408de5"], "compiled_tokens_by_order": [302, 302, 302, 302, 302, 302], "choice": "darwin", "probs": {"darwin": 0.8637954979616888, "einstein": 0.005820208283314497, "hypatia": 0.0230193703038867, "ibn_sina": 0.012321381032466019, "sagan": 0.09104337566519145, "socrates": 0.004000166753452463}, "confidence": 0.6999151619930629, "seconds": 0.17547458700028074, "seconds_all_orders": 1.0528475220016844, "seconds_by_order": [0.1759140110007138, 0.1758072609991359, 0.174934098002268, 0.1750296669997624, 0.17528447500080802, 0.17587800999899628], "argmax_stable": true, "spread": {"darwin": 0.13680068063396145, "einstein": 0.051984785033144944, "hypatia": 0.017797765501876346, "ibn_sina": 0.007475034729177591, "sagan": 0.07976621621593218, "socrates": 0.012160694532081216}, "spread_2dp": {"darwin": 0.13, "einstein": 0.049999999999999996, "hypatia": 0.01, "ibn_sina": 0.0, "sagan": 0.08, "socrates": 0.01}, "choices_by_order": ["darwin", "darwin", "darwin", "darwin", "darwin", "darwin"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.8637954979616888, "einstein": 0.005820208283314497, "hypatia": 0.0230193703038867, "ibn_sina": 0.012321381032466019, "sagan": 0.09104337566519145, "socrates": 0.004000166753452463}, {"ibn_sina": 0.012525978754402601, "hypatia": 0.005221604802010355, "sagan": 0.028227728866939836, "socrates": 0.005916853403924347, "darwin": 0.9347739991961385, "einstein": 0.013333834976584401}, {"hypatia": 0.007102900165135986, "darwin": 0.9302963344696694, "sagan": 0.040874368297693094, "ibn_sina": 0.007560997868034006, "einstein": 0.011710702585919153, "socrates": 0.002454696613548462}, {"ibn_sina": 0.009561392885386693, "einstein": 0.007446420266392515, "darwin": 0.8606891587797291, "sagan": 0.10279466098804899, "hypatia": 0.015764071827657912, "socrates": 0.003744295252784943}, {"einstein": 0.03488736555258924, "sagan": 0.03953256430749339, "socrates": 0.012056750916747332, "ibn_sina": 0.007312789087525657, "darwin": 0.8997570164166485, "hypatia": 0.006453513718995822}, {"socrates": 0.014615391145629677, "ibn_sina": 0.00505094402522501, "hypatia": 0.01656140786763686, "einstein": 0.05780499331645944, "sagan": 0.10799394508287201, "darwin": 0.797973318562177}], "prompt_tokens": 302, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.8637954979616888, "einstein": 0.005820208283314497, "hypatia": 0.0230193703038867, "ibn_sina": 0.012321381032466019, "sagan": 0.09104337566519145, "socrates": 0.004000166753452463}, "confidence": 0.6999151619930629, "choice": "darwin"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 175.8887309988495, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 302, "compiled_prompt_sha256": "3824b179b4488ce489b4ef53aaa0bc5aac8823a7a8d88bcb5a7f0cf96ad47e1b", "compiled_ids_sha256": "4010bc1586d12cb94c392181608c684fece802cc25b6f26e7c35ffddfc2a80eb", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.375, 15.375, 16.75, 16.125, 18.125, 15.0], "runtime_probabilities": [0.8637954592704773, 0.005820208229124546, 0.023019369691610336, 0.012321379967033863, 0.0910433679819107, 0.004000166896730661], "decide_seconds": 0.17213161000108812, "probabilities_f64_minus_runtime_max_abs": 3.8691211456942654e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.012525978754402601, "hypatia": 0.005221604802010355, "sagan": 0.028227728866939836, "socrates": 0.005916853403924347, "darwin": 0.9347739991961385, "einstein": 0.013333834976584401}, "confidence": 0.8136045741256588, "choice": "darwin"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 175.7846519976738, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 302, "compiled_prompt_sha256": "dbc05fef85bcc0d56f0a06133eb4389f636ed6460a3f4b3cf3d4405cbf252d2a", "compiled_ids_sha256": "30493d5211dd4b74c6a5dff1ceb8a027ea94393458146b98e541f5c660849df8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 15.0625, 16.75, 15.1875, 20.25, 16.0], "runtime_probabilities": [0.012525979429483414, 0.00522160530090332, 0.028227729722857475, 0.0059168534353375435, 0.9347740411758423, 0.013333834707736969], "decide_seconds": 0.17202170000018668, "probabilities_f64_minus_runtime_max_abs": 4.197970382424643e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.007102900165135986, "darwin": 0.9302963344696694, "sagan": 0.040874368297693094, "ibn_sina": 0.007560997868034006, "einstein": 0.011710702585919153, "socrates": 0.002454696613548462}, "confidence": 0.8120241379544542, "choice": "darwin"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 174.91146799875423, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 302, "compiled_prompt_sha256": "e452353fa315daca4512110c64c8bab66fbc44387cf2c7967e7b40996297fed1", "compiled_ids_sha256": "e6f4c5e9a68bd604e43ed9d7e0f62d833827757f30e8b18d03ceb075daf7654b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.625, 20.5, 17.375, 15.6875, 16.125, 14.5625], "runtime_probabilities": [0.007102899253368378, 0.9302962422370911, 0.04087436571717262, 0.007560997735708952, 0.011710701510310173, 0.0024546964559704065], "decide_seconds": 0.1712733560016204, "probabilities_f64_minus_runtime_max_abs": 9.223257835255794e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.009561392885386693, "einstein": 0.007446420266392515, "darwin": 0.8606891587797291, "sagan": 0.10279466098804899, "hypatia": 0.015764071827657912, "socrates": 0.003744295252784943}, "confidence": 0.7040489061575741, "choice": "darwin"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 175.0071679998655, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 302, "compiled_prompt_sha256": "c668d5cb4b9609114af6406ec50b06bc347bf14cd986105eb07aab8d3ce32c19", "compiled_ids_sha256": "ca2d8c3ec6069b674966e5365e7616dc00910813aeb55ee7c22d1291bfec4663", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.625, 15.375, 20.125, 18.0, 16.125, 14.6875], "runtime_probabilities": [0.009561393409967422, 0.007446420378983021, 0.8606891632080078, 0.10279466211795807, 0.01576407253742218, 0.003744295332580805], "decide_seconds": 0.1713899450005556, "probabilities_f64_minus_runtime_max_abs": 4.428278743695557e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.03488736555258924, "sagan": 0.03953256430749339, "socrates": 0.012056750916747332, "ibn_sina": 0.007312789087525657, "darwin": 0.8997570164166485, "hypatia": 0.006453513718995822}, "confidence": 0.7423730678682114, "choice": "darwin"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 175.261626001884, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 302, "compiled_prompt_sha256": "224425416d064d538ee268e1d83ab285dc072fbc957d6598f7ce41c93e35e9cf", "compiled_ids_sha256": "26ba3d136f4edd057f6232d7696f6db1b956446a066809894f399bd60521bd41", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 17.0, 15.8125, 15.3125, 20.125, 15.1875], "runtime_probabilities": [0.03488736227154732, 0.03953256458044052, 0.012056751176714897, 0.007312789559364319, 0.8997569680213928, 0.006453513167798519], "decide_seconds": 0.17160853399764164, "probabilities_f64_minus_runtime_max_abs": 4.839525569177994e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.014615391145629677, "ibn_sina": 0.00505094402522501, "hypatia": 0.01656140786763686, "einstein": 0.05780499331645944, "sagan": 0.10799394508287201, "darwin": 0.797973318562177}, "confidence": 0.5860976029247277, "choice": "darwin"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 175.85636999865528, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 302, "compiled_prompt_sha256": "d761f71fff8dabb20c919c88489e7ab111f0c5dc12bd69b7aa2258f047408de5", "compiled_ids_sha256": "ecb3959f6e14b31a8b1e7778935fab75c3e03de4dd2f588aeb0cb8a15c1cdea9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 14.9375, 16.125, 17.375, 18.0, 20.0], "runtime_probabilities": [0.014615391381084919, 0.005050943698734045, 0.01656140759587288, 0.05780499428510666, 0.10799393802881241, 0.7979732751846313], "decide_seconds": 0.17219490900242818, "probabilities_f64_minus_runtime_max_abs": 4.3377545666523076e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:13Z"} {"id": "c-sagan-01-starter-03", "uid": "c089", "task": "kperm", "label": "sagan", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "prompt_sha256_by_order": ["d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f"], "request_sha256": "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "compiled_prompt_sha256_by_order": ["ce02283a565951bd24fed4ecdd1be1ff580fd4e71e1ed42c0be57924ef7735f0", "ba4e9a9fb6557bd03d3c8dfc7d34437d16c0f3251f9ba2798b92a02599923bae", "137e9c0c7561ce673df16cf8f93803152e90d488fcb4915a3b14984231405acf", "52dedeb334558076d3d1b6cc3cd0492d986dda37e5fe9cdd5907369c6f87e4f3", "b77101de530c3776c0c34577a0fa1c2fc7bf98cdf89c29c7fa5752c979c8777a", "334b36f7502e98d9cde994d77acd7496f6f4d6800dec36889109609a6a234c98"], "compiled_tokens_by_order": [355, 355, 355, 355, 355, 355], "choice": "sagan", "probs": {"darwin": 0.011314368459757454, "einstein": 0.330654314673406, "hypatia": 0.01646230897077406, "ibn_sina": 0.021138023135840504, "sagan": 0.6177435866961126, "socrates": 0.002687398064109408}, "confidence": 0.5092936989772234, "seconds": 0.1874200674998671, "seconds_all_orders": 1.1245204049992026, "seconds_by_order": [0.18762764900020557, 0.1880284959988785, 0.18731451100029517, 0.18754176999937044, 0.1867871960021148, 0.1872207829983381], "argmax_stable": false, "spread": {"darwin": 0.06611428656246955, "einstein": 0.5403234316088358, "hypatia": 0.04364967496667599, "ibn_sina": 0.23034143183848327, "sagan": 0.5171293174214899, "socrates": 0.07052972355707166}, "spread_2dp": {"darwin": 0.060000000000000005, "einstein": 0.5399999999999999, "hypatia": 0.049999999999999996, "ibn_sina": 0.23, "sagan": 0.52, "socrates": 0.07}, "choices_by_order": ["sagan", "einstein", "einstein", "sagan", "einstein", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.011314368459757454, "einstein": 0.330654314673406, "hypatia": 0.01646230897077406, "ibn_sina": 0.021138023135840504, "sagan": 0.6177435866961126, "socrates": 0.002687398064109408}, {"ibn_sina": 0.04421382100968863, "hypatia": 0.012667471806793647, "sagan": 0.41948905405596437, "socrates": 0.009267720353091187, "darwin": 0.03901856009247993, "einstein": 0.4753433726819823}, {"hypatia": 0.02347387571020101, "darwin": 0.008112354372586731, "sagan": 0.2227139317939001, "ibn_sina": 0.05631090049485558, "einstein": 0.6860072052003523, "socrates": 0.0033817324281042835}, {"ibn_sina": 0.06073003209374611, "einstein": 0.1456837735915165, "darwin": 0.01971615476198644, "sagan": 0.73984324921539, "hypatia": 0.032506443732502666, "socrates": 0.0015203466048582923}, {"einstein": 0.3229060119618122, "sagan": 0.2514794549743238, "socrates": 0.07205007016192995, "ibn_sina": 0.2514794549743238, "darwin": 0.07205007016192995, "hypatia": 0.030034937765680357}, {"socrates": 0.02660233648159361, "ibn_sina": 0.023476479546519543, "hypatia": 0.056317146773469634, "einstein": 0.4161305568408784, "sagan": 0.4715376967580784, "darwin": 0.005935783599460399}], "prompt_tokens": 355, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.011314368459757454, "einstein": 0.330654314673406, "hypatia": 0.01646230897077406, "ibn_sina": 0.021138023135840504, "sagan": 0.6177435866961126, "socrates": 0.002687398064109408}, "confidence": 0.5092936989772234, "choice": "sagan"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.59866900290945, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "ce02283a565951bd24fed4ecdd1be1ff580fd4e71e1ed42c0be57924ef7735f0", "compiled_ids_sha256": "fcc59376315f837186e1531a6bc1a77307e5878970fb6cc3e97bcfa34479b565", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.625, 19.0, 16.0, 16.25, 19.625, 14.1875], "runtime_probabilities": [0.011314368806779385, 0.3306543231010437, 0.016462307423353195, 0.02113802172243595, 0.6177435517311096, 0.002687398111447692], "decide_seconds": 0.18329658299990115, "probabilities_f64_minus_runtime_max_abs": 3.4965003004039374e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.04421382100968863, "hypatia": 0.012667471806793647, "sagan": 0.41948905405596437, "socrates": 0.009267720353091187, "darwin": 0.03901856009247993, "einstein": 0.4753433726819823}, "confidence": 0.396615753457343, "choice": "einstein"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 188.00267600090592, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "ba4e9a9fb6557bd03d3c8dfc7d34437d16c0f3251f9ba2798b92a02599923bae", "compiled_ids_sha256": "1f5cada7fea9a22086a3299315822a482ee78493540d6047321546300c4ed697", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 15.75, 19.25, 15.4375, 16.875, 19.375], "runtime_probabilities": [0.04421381652355194, 0.01266747061163187, 0.4194889962673187, 0.009267720393836498, 0.03901855647563934, 0.4753433167934418], "decide_seconds": 0.18364323000059812, "probabilities_f64_minus_runtime_max_abs": 5.7788645646539294e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.02347387571020101, "darwin": 0.008112354372586731, "sagan": 0.2227139317939001, "ibn_sina": 0.05631090049485558, "einstein": 0.6860072052003523, "socrates": 0.0033817324281042835}, "confidence": 0.4969269342987723, "choice": "einstein"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.28771200039773, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "137e9c0c7561ce673df16cf8f93803152e90d488fcb4915a3b14984231405acf", "compiled_ids_sha256": "e3e91a6cf40aacb212487a26c4dfef10188da63f6c7e0c252e48dadc5f4cbba3", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 15.1875, 18.5, 17.125, 19.625, 14.3125], "runtime_probabilities": [0.02347387745976448, 0.008112355135381222, 0.22271394729614258, 0.05631090700626373, 0.6860072612762451, 0.0033817326184362173], "decide_seconds": 0.1831364140016376, "probabilities_f64_minus_runtime_max_abs": 5.6075892818086004e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.06073003209374611, "einstein": 0.1456837735915165, "darwin": 0.01971615476198644, "sagan": 0.73984324921539, "hypatia": 0.032506443732502666, "socrates": 0.0015203466048582923}, "confidence": 0.5131383368120624, "choice": "sagan"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.516770001821, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "52dedeb334558076d3d1b6cc3cd0492d986dda37e5fe9cdd5907369c6f87e4f3", "compiled_ids_sha256": "1eba59e423a50ec4dac2773808a644e078beb3ac9d36000edd522f9d8af4c8ce", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.25, 18.125, 16.125, 19.75, 16.625, 13.5625], "runtime_probabilities": [0.06073003634810448, 0.14568376541137695, 0.019716154783964157, 0.7398433089256287, 0.03250644728541374, 0.0015203466173261404], "decide_seconds": 0.18324777300222195, "probabilities_f64_minus_runtime_max_abs": 5.971023864237424e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.3229060119618122, "sagan": 0.2514794549743238, "socrates": 0.07205007016192995, "ibn_sina": 0.2514794549743238, "darwin": 0.07205007016192995, "hypatia": 0.030034937765680357}, "confidence": 0.1384910059531833, "choice": "einstein"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 186.7610560002504, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "b77101de530c3776c0c34577a0fa1c2fc7bf98cdf89c29c7fa5752c979c8777a", "compiled_ids_sha256": "4b675927e7bc281b123d8a453579b681e3f71b80fb3ce354b839ff51b9fe3b7c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.0, 18.75, 17.5, 18.75, 17.5, 16.625], "runtime_probabilities": [0.3229059875011444, 0.25147944688796997, 0.0720500648021698, 0.25147944688796997, 0.0720500648021698, 0.030034935101866722], "decide_seconds": 0.18253542900129105, "probabilities_f64_minus_runtime_max_abs": 2.446066776906619e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.02660233648159361, "ibn_sina": 0.023476479546519543, "hypatia": 0.056317146773469634, "einstein": 0.4161305568408784, "sagan": 0.4715376967580784, "darwin": 0.005935783599460399}, "confidence": 0.38812842125459357, "choice": "sagan"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.19494299875805, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "334b36f7502e98d9cde994d77acd7496f6f4d6800dec36889109609a6a234c98", "compiled_ids_sha256": "bfb720ffd6b7eb966c39f15507c690820bb76fb08be08c40507efe7ccd9bd5e5", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 16.25, 17.125, 19.125, 19.25, 14.875], "runtime_probabilities": [0.02660233899950981, 0.023476479575037956, 0.056317150592803955, 0.41613057255744934, 0.471537709236145, 0.005935783963650465], "decide_seconds": 0.18298391500138678, "probabilities_f64_minus_runtime_max_abs": 1.5716570933577145e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:14Z"} {"id": "c-socrates-01-dessert-01", "uid": "c090", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "prompt_sha256_by_order": ["b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee"], "request_sha256": "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "compiled_prompt_sha256_by_order": ["e5874d1b8d5df9f9ce009834e203a80ed9bfc65e1e117e65b805719bfafff002", "826cca657c25104c48f94c082d5d94b4889b76d32f854dc1bd230a522523032f", "c604dc6c278592df989e225e1321c8e7d15f9b05fd35a13b0ec2f4697065d631", "3050f73463d83ddbc40ca354099fb692e24640d71af15c37a17ce5bb9d2d6274", "99440821fae71a77847677cf8db901ebc7bdf278b18cebc2be67df818e7af741", "8967dc6aabe1738656ebdcacb7913cbd49fb8eb32772829f8826cceafde8b0f8"], "compiled_tokens_by_order": [409, 409, 409, 409, 409, 409], "choice": "socrates", "probs": {"darwin": 0.04842164538821716, "einstein": 0.013873033675587317, "hypatia": 0.04842164538821716, "ibn_sina": 0.020185144892688738, "sagan": 0.010804329490123374, "socrates": 0.8582942011651663}, "confidence": 0.6587587135835935, "seconds": 0.21381003850001434, "seconds_all_orders": 1.282860231000086, "seconds_by_order": [0.21476410800096346, 0.21384472499994445, 0.21423239200157695, 0.21361245599837275, 0.21332427900051698, 0.21308227099871146], "argmax_stable": true, "spread": {"darwin": 0.045837564084420884, "einstein": 0.016542875353312527, "hypatia": 0.03882061860112747, "ibn_sina": 0.03090616000313128, "sagan": 0.010561078372459618, "socrates": 0.12103715555926453}, "spread_2dp": {"darwin": 0.05, "einstein": 0.02, "hypatia": 0.04, "ibn_sina": 0.03, "sagan": 0.01, "socrates": 0.12}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.04842164538821716, "einstein": 0.013873033675587317, "hypatia": 0.04842164538821716, "ibn_sina": 0.020185144892688738, "sagan": 0.010804329490123374, "socrates": 0.8582942011651663}, {"ibn_sina": 0.016851593777443736, "hypatia": 0.021637874721934743, "sagan": 0.007477849409319721, "socrates": 0.920065845358473, "darwin": 0.019095357420620927, "einstein": 0.014871479312207938}, {"hypatia": 0.04465421207108292, "darwin": 0.018614645031104056, "sagan": 0.012793645958377462, "ibn_sina": 0.01060630502718591, "einstein": 0.016427366582661057, "socrates": 0.8969038253295886}, {"ibn_sina": 0.03365690023241565, "einstein": 0.018015240497923937, "darwin": 0.03365690023241565, "sagan": 0.008509797050804652, "hypatia": 0.038138264433386304, "socrates": 0.8680228975530538}, {"einstein": 0.006876792294492393, "sagan": 0.002232567585917843, "socrates": 0.9587710488933849, "ibn_sina": 0.009399460621683809, "darwin": 0.012069146341397571, "hypatia": 0.010650984263123558}, {"socrates": 0.9793313567244308, "ibn_sina": 0.0027507402292843756, "hypatia": 0.009601026787089689, "einstein": 0.0014723651446114094, "sagan": 0.004260429810787443, "darwin": 0.0025840813037962774}], "prompt_tokens": 409, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.04842164538821716, "einstein": 0.013873033675587317, "hypatia": 0.04842164538821716, "ibn_sina": 0.020185144892688738, "sagan": 0.010804329490123374, "socrates": 0.8582942011651663}, "confidence": 0.6587587135835935, "choice": "socrates"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 214.73615800277912, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "e5874d1b8d5df9f9ce009834e203a80ed9bfc65e1e117e65b805719bfafff002", "compiled_ids_sha256": "2fbf7e93347627cacd15188db8ccc02538cff99269ff3adb9d0ea548870e9cb9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 16.125, 17.375, 16.5, 15.875, 20.25], "runtime_probabilities": [0.048421647399663925, 0.013873034156858921, 0.048421647399663925, 0.02018514648079872, 0.010804329998791218, 0.8582941889762878], "decide_seconds": 0.20987382599923876, "probabilities_f64_minus_runtime_max_abs": 1.2188878462104924e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.016851593777443736, "hypatia": 0.021637874721934743, "sagan": 0.007477849409319721, "socrates": 0.920065845358473, "darwin": 0.019095357420620927, "einstein": 0.014871479312207938}, "confidence": 0.7749781903487486, "choice": "socrates"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 213.81907500108355, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "826cca657c25104c48f94c082d5d94b4889b76d32f854dc1bd230a522523032f", "compiled_ids_sha256": "193f094f387d5b91ff8d61100935c73c288a7b16a6fe414732858b803e68dfc7", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 16.625, 15.5625, 20.375, 16.5, 16.25], "runtime_probabilities": [0.016851594671607018, 0.021637877449393272, 0.007477850653231144, 0.9200658798217773, 0.01909535750746727, 0.014871479943394661], "decide_seconds": 0.2089678529991943, "probabilities_f64_minus_runtime_max_abs": 3.446330432144151e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.04465421207108292, "darwin": 0.018614645031104056, "sagan": 0.012793645958377462, "ibn_sina": 0.01060630502718591, "einstein": 0.016427366582661057, "socrates": 0.8969038253295886}, "confidence": 0.7309633491493932, "choice": "socrates"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 214.206713000749, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "c604dc6c278592df989e225e1321c8e7d15f9b05fd35a13b0ec2f4697065d631", "compiled_ids_sha256": "cae13c3737b774dd07b3d009ba5fc432b7b06ce6538c5c9ce26494ea03bb8699", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 16.5, 16.125, 15.9375, 16.375, 20.375], "runtime_probabilities": [0.04465421289205551, 0.018614644184708595, 0.012793646194040775, 0.010606304742395878, 0.01642736792564392, 0.8969038128852844], "decide_seconds": 0.20939871899827267, "probabilities_f64_minus_runtime_max_abs": 1.244430414271136e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.03365690023241565, "einstein": 0.018015240497923937, "darwin": 0.03365690023241565, "sagan": 0.008509797050804652, "hypatia": 0.038138264433386304, "socrates": 0.8680228975530538}, "confidence": 0.6714648415345732, "choice": "socrates"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 213.58788699944853, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "3050f73463d83ddbc40ca354099fb692e24640d71af15c37a17ce5bb9d2d6274", "compiled_ids_sha256": "b8d5c80860c4fcf8751da213e0ce5b2ccf8cde7751880e1e472956126126543c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 16.5, 17.125, 15.75, 17.25, 20.375], "runtime_probabilities": [0.03365689888596535, 0.018015239387750626, 0.03365689888596535, 0.008509797044098377, 0.038138266652822495, 0.8680229187011719], "decide_seconds": 0.2088735639990773, "probabilities_f64_minus_runtime_max_abs": 2.1148118034552965e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.006876792294492393, "sagan": 0.002232567585917843, "socrates": 0.9587710488933849, "ibn_sina": 0.009399460621683809, "darwin": 0.012069146341397571, "hypatia": 0.010650984263123558}, "confidence": 0.8695156750745034, "choice": "socrates"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 213.2992490005563, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "99440821fae71a77847677cf8db901ebc7bdf278b18cebc2be67df818e7af741", "compiled_ids_sha256": "4f9249d1be322fdd961969450b842cde09ad594b0ff5d94e319d6d7b6fc23857", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.8125, 14.6875, 20.75, 16.125, 16.375, 16.25], "runtime_probabilities": [0.006876791827380657, 0.0022325674071907997, 0.9587710499763489, 0.00939946062862873, 0.012069146148860455, 0.010650984942913055], "decide_seconds": 0.2085754559993802, "probabilities_f64_minus_runtime_max_abs": 1.08296394074614e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.9793313567244308, "ibn_sina": 0.0027507402292843756, "hypatia": 0.009601026787089689, "einstein": 0.0014723651446114094, "sagan": 0.004260429810787443, "darwin": 0.0025840813037962774}, "confidence": 0.9277079192387969, "choice": "socrates"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 213.05703100006212, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 409, "compiled_prompt_sha256": "8967dc6aabe1738656ebdcacb7913cbd49fb8eb32772829f8826cceafde8b0f8", "compiled_ids_sha256": "32c12ebbe401fa75f69eef505c22f6e7bd7769a0a47e74220ed78170cf127da8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.875, 15.0, 16.25, 14.375, 15.4375, 14.9375], "runtime_probabilities": [0.979331374168396, 0.0027507406193763018, 0.009601027704775333, 0.0014723651111125946, 0.004260430112481117, 0.0025840813759714365], "decide_seconds": 0.20830510800078628, "probabilities_f64_minus_runtime_max_abs": 1.7443965161234587e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:15Z"} {"id": "c-socrates-01-dessert-01", "uid": "c091", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "prompt_sha256_by_order": ["42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512"], "request_sha256": "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "compiled_prompt_sha256_by_order": ["38c5218cbd49d0ad53554b844ea9581c36f8d9f6dab540ad7633c7aefaaec7f9", "8f189933761c184a074abc08bc8bdc517ccc27b4dd31089bffe4818595ededd7", "3828f5a05800bfa57043bb88653ff9a2be72492389e3b064f30b37f499644d4b", "95574326175c1fb3153918b7d4aefc7008232a67ab2aa789172cef901c60d16e", "9e178a50169ff30850f5576659d9957fd1fad665dfdbf39606bb52e71b42bc9a", "70a83edfe0e32d7f12da55e8b50a5b267829b2bde52579d104e353a309346941"], "compiled_tokens_by_order": [291, 291, 291, 291, 291, 291], "choice": "socrates", "probs": {"darwin": 0.015373085726577852, "einstein": 0.004404462803207475, "hypatia": 0.015373085726577852, "ibn_sina": 0.008759320210120164, "sagan": 0.004990910212044929, "socrates": 0.9510991353214717}, "confidence": 0.850481875419052, "seconds": 0.17459696766551738, "seconds_all_orders": 1.0475818059931044, "seconds_by_order": [0.17468863999965834, 0.17528833599862992, 0.17441039299956174, 0.17433754299781867, 0.17425822299992433, 0.17459867099751136], "argmax_stable": true, "spread": {"darwin": 0.01743878241919892, "einstein": 0.00861469751516591, "hypatia": 0.018589648343942403, "ibn_sina": 0.02622387377250645, "sagan": 0.006493175487888709, "socrates": 0.07641137688877941}, "spread_2dp": {"darwin": 0.02, "einstein": 0.01, "hypatia": 0.01, "ibn_sina": 0.03, "sagan": 0.01, "socrates": 0.07999999999999996}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.015373085726577852, "einstein": 0.004404462803207475, "hypatia": 0.015373085726577852, "ibn_sina": 0.008759320210120164, "sagan": 0.004990910212044929, "socrates": 0.9510991353214717}, {"ibn_sina": 0.015521274359214583, "hypatia": 0.005709957737536497, "sagan": 0.0022360515781060177, "socrates": 0.9602672413785139, "darwin": 0.012087980625222436, "einstein": 0.0041774943214065085}, {"hypatia": 0.017539249824412684, "darwin": 0.009388083916166582, "sagan": 0.0032444349858437994, "ibn_sina": 0.005349168972564341, "einstein": 0.0068684689189300666, "socrates": 0.9576105933820825}, {"ibn_sina": 0.027535061041361063, "einstein": 0.009515862467536901, "darwin": 0.018924552244582442, "sagan": 0.007888927070188087, "hypatia": 0.0242996060814789, "socrates": 0.9118359910948527}, {"einstein": 0.0053807318189606325, "sagan": 0.002880098199894453, "socrates": 0.963260988092252, "ibn_sina": 0.01070086298214069, "darwin": 0.00833384047003098, "hypatia": 0.009443478436721315}, {"socrates": 0.9882473679836321, "ibn_sina": 0.0013111872688546132, "hypatia": 0.0066587583874594255, "einstein": 0.0009011649523709917, "sagan": 0.0013957515822993775, "darwin": 0.0014857698253835218}], "prompt_tokens": 291, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.015373085726577852, "einstein": 0.004404462803207475, "hypatia": 0.015373085726577852, "ibn_sina": 0.008759320210120164, "sagan": 0.004990910212044929, "socrates": 0.9510991353214717}, "confidence": 0.850481875419052, "choice": "socrates"}}, "usage": {"input_tokens": 291, "output_tokens": 1}, "latency_ms": 174.66348100060713, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 291, "compiled_prompt_sha256": "38c5218cbd49d0ad53554b844ea9581c36f8d9f6dab540ad7633c7aefaaec7f9", "compiled_ids_sha256": "2b8caa07c5125e82419f0a570ae0c7c4b742c0c93eaa8a8cd7f945e8e8e0643e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 15.25, 16.5, 15.9375, 15.375, 20.625], "runtime_probabilities": [0.015373085625469685, 0.004404463339596987, 0.015373085625469685, 0.00875931978225708, 0.004990910179913044, 0.951099157333374], "decide_seconds": 0.17089951899833977, "probabilities_f64_minus_runtime_max_abs": 2.201190230088912e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.015521274359214583, "hypatia": 0.005709957737536497, "sagan": 0.0022360515781060177, "socrates": 0.9602672413785139, "darwin": 0.012087980625222436, "einstein": 0.0041774943214065085}, "confidence": 0.8755477205002283, "choice": "socrates"}}, "usage": {"input_tokens": 291, "output_tokens": 1}, "latency_ms": 175.26619599811966, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 291, "compiled_prompt_sha256": "8f189933761c184a074abc08bc8bdc517ccc27b4dd31089bffe4818595ededd7", "compiled_ids_sha256": "f93e85791447fcb0c6ec05b6330a6b949d655214d143738ba2cac876e044a327", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 15.375, 14.4375, 20.5, 16.125, 15.0625], "runtime_probabilities": [0.015521273016929626, 0.005709957331418991, 0.0022360514849424362, 0.960267186164856, 0.012087980285286903, 0.00417749397456646], "decide_seconds": 0.1716174939974735, "probabilities_f64_minus_runtime_max_abs": 5.521365797900302e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.017539249824412684, "darwin": 0.009388083916166582, "sagan": 0.0032444349858437994, "ibn_sina": 0.005349168972564341, "einstein": 0.0068684689189300666, "socrates": 0.9576105933820825}, "confidence": 0.8677246822661424, "choice": "socrates"}}, "usage": {"input_tokens": 291, "output_tokens": 1}, "latency_ms": 174.38816299909377, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 291, "compiled_prompt_sha256": "3828f5a05800bfa57043bb88653ff9a2be72492389e3b064f30b37f499644d4b", "compiled_ids_sha256": "8b6092398dd4000f63708de61cb52a18bd7834aee196cd8c7b3f0257c113659b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 16.0, 14.9375, 15.4375, 15.6875, 20.625], "runtime_probabilities": [0.01753925159573555, 0.009388084523379803, 0.0032444351818412542, 0.005349169485270977, 0.006868469063192606, 0.9576106667518616], "decide_seconds": 0.1708735499996692, "probabilities_f64_minus_runtime_max_abs": 7.336977903005248e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.027535061041361063, "einstein": 0.009515862467536901, "darwin": 0.018924552244582442, "sagan": 0.007888927070188087, "hypatia": 0.0242996060814789, "socrates": 0.9118359910948527}, "confidence": 0.7594682218585702, "choice": "socrates"}}, "usage": {"input_tokens": 291, "output_tokens": 1}, "latency_ms": 174.3156230004388, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 291, "compiled_prompt_sha256": "95574326175c1fb3153918b7d4aefc7008232a67ab2aa789172cef901c60d16e", "compiled_ids_sha256": "128d2c2333aeb543374911fd68ac206b72525444319b8b78b7d1593ad2c78666", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 15.8125, 16.5, 15.625, 16.75, 20.375], "runtime_probabilities": [0.027535060420632362, 0.009515861980617046, 0.018924551084637642, 0.007888926193118095, 0.024299604818224907, 0.9118359684944153], "decide_seconds": 0.17067431200121064, "probabilities_f64_minus_runtime_max_abs": 2.2600437410424945e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.0053807318189606325, "sagan": 0.002880098199894453, "socrates": 0.963260988092252, "ibn_sina": 0.01070086298214069, "darwin": 0.00833384047003098, "hypatia": 0.009443478436721315}, "confidence": 0.8808435381999975, "choice": "socrates"}}, "usage": {"input_tokens": 291, "output_tokens": 1}, "latency_ms": 174.2371729997103, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 291, "compiled_prompt_sha256": "9e178a50169ff30850f5576659d9957fd1fad665dfdbf39606bb52e71b42bc9a", "compiled_ids_sha256": "31a94af66f20d8fb8bc8b4405817e0492d805ffff7fa1acef161bbe9042c2243", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.5625, 14.9375, 20.75, 16.25, 16.0, 16.125], "runtime_probabilities": [0.005380732007324696, 0.0028800980653613806, 0.9632609486579895, 0.010700862854719162, 0.008333840407431126, 0.009443478658795357], "decide_seconds": 0.17070930100089754, "probabilities_f64_minus_runtime_max_abs": 3.94342625220645e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.9882473679836321, "ibn_sina": 0.0013111872688546132, "hypatia": 0.0066587583874594255, "einstein": 0.0009011649523709917, "sagan": 0.0013957515822993775, "darwin": 0.0014857698253835218}, "confidence": 0.9559494857660524, "choice": "socrates"}}, "usage": {"input_tokens": 291, "output_tokens": 1}, "latency_ms": 174.57692099924316, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 291, "compiled_prompt_sha256": "70a83edfe0e32d7f12da55e8b50a5b267829b2bde52579d104e353a309346941", "compiled_ids_sha256": "62dd4eda8cbfe25d72fdf9049c9a989998478590190a0c47e033df08a312ecc9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.75, 14.125, 15.75, 13.75, 14.1875, 14.25], "runtime_probabilities": [0.9882473349571228, 0.0013111871667206287, 0.006658758036792278, 0.0009011649526655674, 0.0013957516057416797, 0.001485769753344357], "decide_seconds": 0.17104059799748939, "probabilities_f64_minus_runtime_max_abs": 3.302650930514517e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:16Z"} {"id": "c-socrates-01-dessert-01", "uid": "c092", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "prompt_sha256_by_order": ["67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261"], "request_sha256": "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "compiled_prompt_sha256_by_order": ["9a53e19f51ada7e3ca7748e01853f6e0dfe0321fc114aca34ef2b3bce186324a", "de8d7cb4024a379a66d860fe4bf7cdef4bcd513d536e55204b94c41028771cc6", "51f75317bdb836de7c8910f78a3432898acf6b93f46c3428c37789edc7151bf3", "f84157b0f20df3bff51dcaa2026f36e96c6e497703ac65f24ff3bfd71fedcffc", "7ba923ff5a29def4080b4cc3b74adbc94ad27b95cd6f272861e5486c94febab7", "cddcd9cbdf38b13843bf7a9aaff4de5fd4e017ae4dcc0dae94ea85c961efcdc9"], "compiled_tokens_by_order": [361, 361, 361, 361, 361, 361], "choice": "ibn_sina", "probs": {"darwin": 0.13240502577969998, "einstein": 0.13240502577969998, "hypatia": 0.13240502577969998, "ibn_sina": 0.40783619128805815, "sagan": 0.13240502577969998, "socrates": 0.06254370559314205}, "confidence": 0.10145208171948494, "seconds": 0.19299589416732488, "seconds_all_orders": 1.1579753650039493, "seconds_by_order": [0.19293168800140847, 0.1929211979986576, 0.19332143499923404, 0.1930584670008102, 0.19283930900201085, 0.1929032680018281], "argmax_stable": false, "spread": {"darwin": 0.28780583589625475, "einstein": 0.33350754601892446, "hypatia": 0.07815149348104075, "ibn_sina": 0.36356675720865345, "sagan": 0.13229310060039226, "socrates": 0.613828119986001}, "spread_2dp": {"darwin": 0.29000000000000004, "einstein": 0.33, "hypatia": 0.08000000000000002, "ibn_sina": 0.36, "sagan": 0.13, "socrates": 0.6100000000000001}, "choices_by_order": ["ibn_sina", "einstein", "einstein", "darwin", "ibn_sina", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.13240502577969998, "einstein": 0.13240502577969998, "hypatia": 0.13240502577969998, "ibn_sina": 0.40783619128805815, "sagan": 0.13240502577969998, "socrates": 0.06254370559314205}, {"ibn_sina": 0.19597845385118276, "hypatia": 0.09257366665719724, "sagan": 0.11886694090381976, "socrates": 0.11886694090381976, "darwin": 0.22207268181589618, "einstein": 0.25164131586808425}, {"hypatia": 0.12317538956105467, "darwin": 0.05134714166710831, "sagan": 0.06593103497483324, "ibn_sina": 0.2954823988427868, "einstein": 0.3794069102980028, "socrates": 0.0846571246562143}, {"ibn_sina": 0.2330962053493346, "einstein": 0.04589936427907829, "darwin": 0.33915297756336305, "sagan": 0.18153550725703477, "hypatia": 0.14137999520704336, "socrates": 0.05893595034414594}, {"einstein": 0.17187288728248426, "sagan": 0.04924240665664251, "socrates": 0.15167729066506366, "ibn_sina": 0.4123016230047433, "darwin": 0.15167729066506366, "hypatia": 0.06322850172600261}, {"socrates": 0.6727640703301468, "ibn_sina": 0.0487348657960899, "hypatia": 0.08035020986272956, "einstein": 0.05522383778725866, "sagan": 0.08035020986272956, "darwin": 0.0625768063610455}], "prompt_tokens": 361, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.13240502577969998, "einstein": 0.13240502577969998, "hypatia": 0.13240502577969998, "ibn_sina": 0.40783619128805815, "sagan": 0.13240502577969998, "socrates": 0.06254370559314205}, "confidence": 0.10145208171948494, "choice": "ibn_sina"}}, "usage": {"input_tokens": 361, "output_tokens": 1}, "latency_ms": 192.9042979973019, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 361, "compiled_prompt_sha256": "9a53e19f51ada7e3ca7748e01853f6e0dfe0321fc114aca34ef2b3bce186324a", "compiled_ids_sha256": "08a973dfb69a66733142afb8c93b88554f14300d2455d7ab0777f84c253ecdf1", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.25, 18.25, 18.25, 19.375, 18.25, 17.5], "runtime_probabilities": [0.13240502774715424, 0.13240502774715424, 0.13240502774715424, 0.4078361988067627, 0.13240502774715424, 0.06254370510578156], "decide_seconds": 0.18849021300047752, "probabilities_f64_minus_runtime_max_abs": 7.518704547848643e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.19597845385118276, "hypatia": 0.09257366665719724, "sagan": 0.11886694090381976, "socrates": 0.11886694090381976, "darwin": 0.22207268181589618, "einstein": 0.25164131586808425}, "confidence": 0.03593179693259341, "choice": "einstein"}}, "usage": {"input_tokens": 361, "output_tokens": 1}, "latency_ms": 192.8972480018274, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 361, "compiled_prompt_sha256": "de8d7cb4024a379a66d860fe4bf7cdef4bcd513d536e55204b94c41028771cc6", "compiled_ids_sha256": "aa9683662fd38d0c550a7d7ecc657147256be25d584da37e731d6af942259ddf", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.625, 17.875, 18.125, 18.125, 18.75, 18.875], "runtime_probabilities": [0.19597844779491425, 0.09257367253303528, 0.1188669428229332, 0.1188669428229332, 0.22207267582416534, 0.25164130330085754], "decide_seconds": 0.188659880997875, "probabilities_f64_minus_runtime_max_abs": 1.2567226703730228e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.12317538956105467, "darwin": 0.05134714166710831, "sagan": 0.06593103497483324, "ibn_sina": 0.2954823988427868, "einstein": 0.3794069102980028, "socrates": 0.0846571246562143}, "confidence": 0.14796141524347795, "choice": "einstein"}}, "usage": {"input_tokens": 361, "output_tokens": 1}, "latency_ms": 193.28809500075295, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 361, "compiled_prompt_sha256": "51f75317bdb836de7c8910f78a3432898acf6b93f46c3428c37789edc7151bf3", "compiled_ids_sha256": "de468230ede7b10401815cdcd6e89ca384dfc7693725852a4e468fc3779c9fe6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.25, 17.375, 17.625, 19.125, 19.375, 17.875], "runtime_probabilities": [0.12317539006471634, 0.05134714022278786, 0.06593102961778641, 0.2954823970794678, 0.3794069290161133, 0.08465712517499924], "decide_seconds": 0.1886866410022776, "probabilities_f64_minus_runtime_max_abs": 1.87181105060219e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.2330962053493346, "einstein": 0.04589936427907829, "darwin": 0.33915297756336305, "sagan": 0.18153550725703477, "hypatia": 0.14137999520704336, "socrates": 0.05893595034414594}, "confidence": 0.10656595031783545, "choice": "darwin"}}, "usage": {"input_tokens": 361, "output_tokens": 1}, "latency_ms": 193.03392700021504, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 361, "compiled_prompt_sha256": "f84157b0f20df3bff51dcaa2026f36e96c6e497703ac65f24ff3bfd71fedcffc", "compiled_ids_sha256": "e79370f98d912912deab76fbcb1ee8c1917dfa2ed819315b041773760006fcb4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.75, 17.125, 19.125, 18.5, 18.25, 17.375], "runtime_probabilities": [0.23309622704982758, 0.04589936509728432, 0.339152991771698, 0.1815355271100998, 0.14137999713420868, 0.058935947716236115], "decide_seconds": 0.18870245099969907, "probabilities_f64_minus_runtime_max_abs": 2.170049298588772e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.17187288728248426, "sagan": 0.04924240665664251, "socrates": 0.15167729066506366, "ibn_sina": 0.4123016230047433, "darwin": 0.15167729066506366, "hypatia": 0.06322850172600261}, "confidence": 0.12770824307753192, "choice": "ibn_sina"}}, "usage": {"input_tokens": 361, "output_tokens": 1}, "latency_ms": 192.81530900116195, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 361, "compiled_prompt_sha256": "7ba923ff5a29def4080b4cc3b74adbc94ad27b95cd6f272861e5486c94febab7", "compiled_ids_sha256": "6ed54a7b4bb717c0b75f1350027b93775f9e1466cf75e1c5351cc67606bc2db4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.375, 17.125, 18.25, 19.25, 18.25, 17.375], "runtime_probabilities": [0.17187288403511047, 0.04924240708351135, 0.15167731046676636, 0.412301629781723, 0.15167731046676636, 0.06322850286960602], "decide_seconds": 0.18861501100036548, "probabilities_f64_minus_runtime_max_abs": 1.98017026942221e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.6727640703301468, "ibn_sina": 0.0487348657960899, "hypatia": 0.08035020986272956, "einstein": 0.05522383778725866, "sagan": 0.08035020986272956, "darwin": 0.0625768063610455}, "confidence": 0.35680122307473405, "choice": "socrates"}}, "usage": {"input_tokens": 361, "output_tokens": 1}, "latency_ms": 192.87995799822966, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 361, "compiled_prompt_sha256": "cddcd9cbdf38b13843bf7a9aaff4de5fd4e017ae4dcc0dae94ea85c961efcdc9", "compiled_ids_sha256": "c9b9f8ec626de520409705fd3ee7a2764748b93fd4a1736c06e601c13f97ab99", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.0, 17.375, 17.875, 17.5, 17.875, 17.625], "runtime_probabilities": [0.6727640628814697, 0.0487348698079586, 0.08035021275281906, 0.05522384122014046, 0.08035021275281906, 0.06257680803537369], "decide_seconds": 0.1886822909982584, "probabilities_f64_minus_runtime_max_abs": 7.448677119548108e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:17Z"} {"id": "c-socrates-01-dessert-02", "uid": "c093", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "prompt_sha256_by_order": ["14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e"], "request_sha256": "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "compiled_prompt_sha256_by_order": ["98bf941e9d49195d9f221459067344330d3de649eeabf47b96667525ca3aa283", "a6bdb550b05093a7263f98cc854cd521f95703cdc49a66b59cea5fe8b2d41ad9", "9bca8f1e371997ae4d4ee49bce13b6556ba34f80e746c6c72e2117d9fc507c3a", "feef5a6a106e90235b1f563bc78fdb009e737e32d309db116fc776e8272ecd4c", "61380b6f56927125bd98ea094e716be7e591d0ff8fbe658560560b004d22de75", "e59ecb5de50a329c375c493f61934017ab12c03525749ada6d1185884cb8f983"], "compiled_tokens_by_order": [343, 343, 343, 343, 343, 343], "choice": "hypatia", "probs": {"darwin": 0.019894858233889274, "einstein": 0.0164934167779653, "hypatia": 0.45280488630390625, "ibn_sina": 0.2746400464110122, "sagan": 0.14700422357048018, "socrates": 0.0891625687027468}, "confidence": 0.24281757229900214, "seconds": 0.186486692832356, "seconds_all_orders": 1.118920156994136, "seconds_by_order": [0.1865550769980473, 0.18655601799764554, 0.18666701599795488, 0.18626076999862562, 0.1863237690013193, 0.1865575070005434], "argmax_stable": false, "spread": {"darwin": 0.05713178626204059, "einstein": 0.21784399557024642, "hypatia": 0.39281647525311336, "ibn_sina": 0.3113725280680236, "sagan": 0.46265599836133275, "socrates": 0.3774755952942014}, "spread_2dp": {"darwin": 0.06, "einstein": 0.21000000000000002, "hypatia": 0.39, "ibn_sina": 0.31, "sagan": 0.46, "socrates": 0.37}, "choices_by_order": ["hypatia", "sagan", "sagan", "sagan", "ibn_sina", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.019894858233889274, "einstein": 0.0164934167779653, "hypatia": 0.45280488630390625, "ibn_sina": 0.2746400464110122, "sagan": 0.14700422357048018, "socrates": 0.0891625687027468}, {"ibn_sina": 0.07702664449592986, "hypatia": 0.059988411050792896, "sagan": 0.5022768162039296, "socrates": 0.09890416929494487, "darwin": 0.07702664449592986, "einstein": 0.1847773144584729}, {"hypatia": 0.09768636699771097, "darwin": 0.02179677471251754, "sagan": 0.4377999232946743, "ibn_sina": 0.09768636699771097, "einstein": 0.23433741234821173, "socrates": 0.11069315564917445}, {"ibn_sina": 0.09757511914429216, "einstein": 0.05918230138612634, "darwin": 0.07599157919786036, "sagan": 0.34057062992878023, "hypatia": 0.34057062992878023, "socrates": 0.08610974041416068}, {"einstein": 0.09504540104937632, "sagan": 0.039620817842596835, "socrates": 0.2583601865511219, "ibn_sina": 0.331741046191827, "darwin": 0.07402143276459, "hypatia": 0.201211115600488}, {"socrates": 0.4635853357083621, "ibn_sina": 0.02036851812380335, "hypatia": 0.17054351423566774, "einstein": 0.10343987019907044, "sagan": 0.21898220692984502, "darwin": 0.02308055480325138}], "prompt_tokens": 343, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.019894858233889274, "einstein": 0.0164934167779653, "hypatia": 0.45280488630390625, "ibn_sina": 0.2746400464110122, "sagan": 0.14700422357048018, "socrates": 0.0891625687027468}, "confidence": 0.24281757229900214, "choice": "hypatia"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 186.52772700079367, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 343, "compiled_prompt_sha256": "98bf941e9d49195d9f221459067344330d3de649eeabf47b96667525ca3aa283", "compiled_ids_sha256": "0f0f1ae75907c024b2d459061eabb496925526ca0d95c76d438b65fd51aa1e6e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.0, 15.8125, 19.125, 18.625, 18.0, 17.5], "runtime_probabilities": [0.019894858822226524, 0.01649341732263565, 0.45280489325523376, 0.2746400535106659, 0.14700421690940857, 0.08916256576776505], "decide_seconds": 0.18220206199839595, "probabilities_f64_minus_runtime_max_abs": 7.09965369827259e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.07702664449592986, "hypatia": 0.059988411050792896, "sagan": 0.5022768162039296, "socrates": 0.09890416929494487, "darwin": 0.07702664449592986, "einstein": 0.1847773144584729}, "confidence": 0.190501967822999, "choice": "sagan"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 186.53212799836183, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 343, "compiled_prompt_sha256": "a6bdb550b05093a7263f98cc854cd521f95703cdc49a66b59cea5fe8b2d41ad9", "compiled_ids_sha256": "60f0d47e1eef9297cc7b2f50cdbff5a6bf7dc1cea5a9ae9f3a1b2dddd8c91f0d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 17.125, 19.25, 17.625, 17.375, 18.25], "runtime_probabilities": [0.07702664285898209, 0.05998840928077698, 0.5022767782211304, 0.09890415519475937, 0.07702664285898209, 0.18477731943130493], "decide_seconds": 0.1824524189978547, "probabilities_f64_minus_runtime_max_abs": 3.798279923916681e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.09768636699771097, "darwin": 0.02179677471251754, "sagan": 0.4377999232946743, "ibn_sina": 0.09768636699771097, "einstein": 0.23433741234821173, "socrates": 0.11069315564917445}, "confidence": 0.17226240941558535, "choice": "sagan"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 186.64381700000376, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 343, "compiled_prompt_sha256": "9bca8f1e371997ae4d4ee49bce13b6556ba34f80e746c6c72e2117d9fc507c3a", "compiled_ids_sha256": "25392c527764f88ca74e3177d6be783bb10f5f379082b62d12bcb17e0d187c0e", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 16.0, 19.0, 17.5, 18.375, 17.625], "runtime_probabilities": [0.09768636524677277, 0.021796774119138718, 0.4377999007701874, 0.09768636524677277, 0.23433741927146912, 0.11069315671920776], "decide_seconds": 0.1826309880016197, "probabilities_f64_minus_runtime_max_abs": 2.2524486942288036e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.09757511914429216, "einstein": 0.05918230138612634, "darwin": 0.07599157919786036, "sagan": 0.34057062992878023, "hypatia": 0.34057062992878023, "socrates": 0.08610974041416068}, "confidence": 0.14326729706279495, "choice": "sagan"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 186.23782999929972, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 343, "compiled_prompt_sha256": "feef5a6a106e90235b1f563bc78fdb009e737e32d309db116fc776e8272ecd4c", "compiled_ids_sha256": "5132c6740c56f42ab95ac4f4a2d9e7a148e2b54be1bcb3e4fda88f8218fd090a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 16.875, 17.125, 18.625, 18.625, 17.25], "runtime_probabilities": [0.0975751206278801, 0.0591822974383831, 0.07599157840013504, 0.3405706286430359, 0.3405706286430359, 0.08610974252223969], "decide_seconds": 0.1821583909986657, "probabilities_f64_minus_runtime_max_abs": 3.947743236476509e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.09504540104937632, "sagan": 0.039620817842596835, "socrates": 0.2583601865511219, "ibn_sina": 0.331741046191827, "darwin": 0.07402143276459, "hypatia": 0.201211115600488}, "confidence": 0.11671774825151104, "choice": "ibn_sina"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 186.30093899992062, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 343, "compiled_prompt_sha256": "61380b6f56927125bd98ea094e716be7e591d0ff8fbe658560560b004d22de75", "compiled_ids_sha256": "ef14bd6259424374d629ed9e6794321d893025b526a5341852bf84399f21a58f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 16.625, 18.5, 18.75, 17.25, 18.25], "runtime_probabilities": [0.09504540264606476, 0.03962082043290138, 0.25836020708084106, 0.331741064786911, 0.07402143627405167, 0.2012111246585846], "decide_seconds": 0.18228718099999242, "probabilities_f64_minus_runtime_max_abs": 2.052971914690005e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.4635853357083621, "ibn_sina": 0.02036851812380335, "hypatia": 0.17054351423566774, "einstein": 0.10343987019907044, "sagan": 0.21898220692984502, "darwin": 0.02308055480325138}, "confidence": 0.22333433614285925, "choice": "socrates"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 186.53296699994826, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 343, "compiled_prompt_sha256": "e59ecb5de50a329c375c493f61934017ab12c03525749ada6d1185884cb8f983", "compiled_ids_sha256": "0c39e6a026de2ad02557d14e5fb1b266a0349c66b025dd19458e82c756828afa", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.25, 16.125, 18.25, 17.75, 18.5, 16.25], "runtime_probabilities": [0.4635853171348572, 0.02036851830780506, 0.17054352164268494, 0.10343986749649048, 0.21898220479488373, 0.023080553859472275], "decide_seconds": 0.18225258099846542, "probabilities_f64_minus_runtime_max_abs": 1.8573504900754045e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:18Z"} {"id": "c-socrates-01-dessert-02", "uid": "c094", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "prompt_sha256_by_order": ["d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6"], "request_sha256": "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "compiled_prompt_sha256_by_order": ["e0d5ffa591269263ed024ca529109dfb26c248d9d18c4a8d6e57f2ba9168b233", "4622674a2ae1102168d7282f9df1a3846c761ba064fef823a6e8d4ebe07ff5f0", "3bd5e7b473e603250aa7e4fad4fa980d5d6ff694a3f53fb5a6faf31bab59e8ad", "8a354122519a3aca753b0519c384a81734212a386aeafce73bd4b02d6bdc04a1", "3c5bdd5f794315afa0e3cd15c25303a46f20d73dfb7a1d058fd2fe4ed4fb7347", "cc2449b92f57f5c6ff39f6eff7855ff197f476dc98099ffbb66a8df50ae2073b"], "compiled_tokens_by_order": [374, 374, 374, 374, 374, 374], "choice": "socrates", "probs": {"darwin": 0.0035644186108922437, "einstein": 0.0010870862567951443, "hypatia": 0.0040390154350151515, "ibn_sina": 0.0016837121317530854, "sagan": 0.0013112763266761272, "socrates": 0.9883144912388682}, "confidence": 0.9548793197358132, "seconds": 0.1952409533332684, "seconds_all_orders": 1.1714457199996104, "seconds_by_order": [0.19531160900078248, 0.1954460079978162, 0.19519008100178326, 0.19531386000016937, 0.1949847920004686, 0.19519936999859056], "argmax_stable": true, "spread": {"darwin": 0.002899256666953673, "einstein": 0.0010818681620315428, "hypatia": 0.019164610531016424, "ibn_sina": 0.0016205614451795922, "sagan": 0.0009336986360880097, "socrates": 0.023982843894909456}, "spread_2dp": {"darwin": 0.0, "einstein": 0.0, "hypatia": 0.02, "ibn_sina": 0.0, "sagan": 0.0, "socrates": 0.030000000000000027}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0035644186108922437, "einstein": 0.0010870862567951443, "hypatia": 0.0040390154350151515, "ibn_sina": 0.0016837121317530854, "sagan": 0.0013112763266761272, "socrates": 0.9883144912388682}, {"ibn_sina": 0.0018008606943109276, "hypatia": 0.001092277224970958, "sagan": 0.00048469508044368914, "socrates": 0.9930338123043752, "darwin": 0.0027892276069680665, "einstein": 0.00079912708893122}, {"hypatia": 0.005877667573819061, "darwin": 0.0023017276936433146, "sagan": 0.0007954558058734569, "ibn_sina": 0.0011573813682540825, "einstein": 0.0013960684165043177, "socrates": 0.9884716991419057}, {"ibn_sina": 0.001999559296500423, "einstein": 0.0011393145441769882, "darwin": 0.002411928908949746, "sagan": 0.0012127940192409306, "hypatia": 0.020194833503286862, "socrates": 0.973041569727845}, {"einstein": 0.00079933197724762, "sagan": 0.0003775776905881175, "socrates": 0.9932884163952006, "ibn_sina": 0.0016921858091122537, "darwin": 0.0018013224172594286, "hypatia": 0.0020411657105921174}, {"socrates": 0.9970244136227544, "ibn_sina": 0.00037899785132083073, "hypatia": 0.001030222972270437, "einstein": 0.0003142002544727747, "sagan": 0.000587003355242937, "darwin": 0.0006651619439385708}], "prompt_tokens": 374, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0035644186108922437, "einstein": 0.0010870862567951443, "hypatia": 0.0040390154350151515, "ibn_sina": 0.0016837121317530854, "sagan": 0.0013112763266761272, "socrates": 0.9883144912388682}, "confidence": 0.9548793197358132, "choice": "socrates"}}, "usage": {"input_tokens": 374, "output_tokens": 1}, "latency_ms": 195.2843489998486, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 374, "compiled_prompt_sha256": "e0d5ffa591269263ed024ca529109dfb26c248d9d18c4a8d6e57f2ba9168b233", "compiled_ids_sha256": "42ef38f2cbdef5189c3d759a7044b4c8a748d18138cc275d8d40d3f3af6771d4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.25, 14.0625, 15.375, 14.5, 14.25, 20.875], "runtime_probabilities": [0.0035644187591969967, 0.0010870862752199173, 0.004039015620946884, 0.0016837123548611999, 0.0013112763408571482, 0.9883145093917847], "decide_seconds": 0.19088342399845715, "probabilities_f64_minus_runtime_max_abs": 1.815291650331119e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0018008606943109276, "hypatia": 0.001092277224970958, "sagan": 0.00048469508044368914, "socrates": 0.9930338123043752, "darwin": 0.0027892276069680665, "einstein": 0.00079912708893122}, "confidence": 0.9712149257981512, "choice": "socrates"}}, "usage": {"input_tokens": 374, "output_tokens": 1}, "latency_ms": 195.4213780009013, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 374, "compiled_prompt_sha256": "4622674a2ae1102168d7282f9df1a3846c761ba064fef823a6e8d4ebe07ff5f0", "compiled_ids_sha256": "392b0391b76bd9c054c06c3726dfba5edfa6caf992952cb90af705f0ebe4c781", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.6875, 14.1875, 13.375, 21.0, 15.125, 13.875], "runtime_probabilities": [0.0018008605111390352, 0.0010922771180048585, 0.0004846950760111213, 0.993033766746521, 0.00278922775760293, 0.0007991270394995809], "decide_seconds": 0.19104704300116282, "probabilities_f64_minus_runtime_max_abs": 4.5557854155831023e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.005877667573819061, "darwin": 0.0023017276936433146, "sagan": 0.0007954558058734569, "ibn_sina": 0.0011573813682540825, "einstein": 0.0013960684165043177, "socrates": 0.9884716991419057}, "confidence": 0.9562920221575517, "choice": "socrates"}}, "usage": {"input_tokens": 374, "output_tokens": 1}, "latency_ms": 195.16629099962302, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 374, "compiled_prompt_sha256": "3bd5e7b473e603250aa7e4fad4fa980d5d6ff694a3f53fb5a6faf31bab59e8ad", "compiled_ids_sha256": "bb3ae6e0798e18997aacf12d2b357fb0ce288f0a4f72d30e266ecea61c5ee1ba", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.75, 14.8125, 13.75, 14.125, 14.3125, 20.875], "runtime_probabilities": [0.00587766757234931, 0.002301727654412389, 0.0007954558241181076, 0.0011573813389986753, 0.0013960684882476926, 0.9884717464447021], "decide_seconds": 0.19090945399875636, "probabilities_f64_minus_runtime_max_abs": 4.730279645759339e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.001999559296500423, "einstein": 0.0011393145441769882, "darwin": 0.002411928908949746, "sagan": 0.0012127940192409306, "hypatia": 0.020194833503286862, "socrates": 0.973041569727845}, "confidence": 0.9172722474184801, "choice": "socrates"}}, "usage": {"input_tokens": 374, "output_tokens": 1}, "latency_ms": 195.28984000135097, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 374, "compiled_prompt_sha256": "8a354122519a3aca753b0519c384a81734212a386aeafce73bd4b02d6bdc04a1", "compiled_ids_sha256": "92b126d4b011e122a037ec61f29c4d0697dd7f29734e6c00331be0ff987b79f4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.6875, 14.125, 14.875, 14.1875, 17.0, 20.875], "runtime_probabilities": [0.001999559113755822, 0.001139314379543066, 0.002411928726360202, 0.0012127939844503999, 0.020194830372929573, 0.9730414748191833], "decide_seconds": 0.1910328319972905, "probabilities_f64_minus_runtime_max_abs": 9.490866159644895e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.00079933197724762, "sagan": 0.0003775776905881175, "socrates": 0.9932884163952006, "ibn_sina": 0.0016921858091122537, "darwin": 0.0018013224172594286, "hypatia": 0.0020411657105921174}, "confidence": 0.971987804051346, "choice": "socrates"}}, "usage": {"input_tokens": 374, "output_tokens": 1}, "latency_ms": 194.96058200093103, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 374, "compiled_prompt_sha256": "3c5bdd5f794315afa0e3cd15c25303a46f20d73dfb7a1d058fd2fe4ed4fb7347", "compiled_ids_sha256": "c06290d6f78e341cb8c609a0f08a06e55061bd9f3aa5755f6d95f0a90134e3f5", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [13.75, 13.0, 20.875, 14.5, 14.5625, 14.6875], "runtime_probabilities": [0.0007993319886736572, 0.0003775777295231819, 0.9932884573936462, 0.0016921859933063388, 0.0018013224471360445, 0.002041165716946125], "decide_seconds": 0.19070627499968396, "probabilities_f64_minus_runtime_max_abs": 4.099844563842936e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.9970244136227544, "ibn_sina": 0.00037899785132083073, "hypatia": 0.001030222972270437, "einstein": 0.0003142002544727747, "sagan": 0.000587003355242937, "darwin": 0.0006651619439385708}, "confidence": 0.9861530035715939, "choice": "socrates"}}, "usage": {"input_tokens": 374, "output_tokens": 1}, "latency_ms": 195.1748999999836, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 374, "compiled_prompt_sha256": "cc2449b92f57f5c6ff39f6eff7855ff197f476dc98099ffbb66a8df50ae2073b", "compiled_ids_sha256": "6bcba28bb55de8670937e51ef0c42644a098379968dfb4426183c8d33b7c67f9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [21.0, 13.125, 14.125, 12.9375, 13.5625, 13.6875], "runtime_probabilities": [0.9970243573188782, 0.00037899785093031824, 0.0010302228620275855, 0.00031420026789419353, 0.00058700330555439, 0.0006651619332842529], "decide_seconds": 0.1907350649998989, "probabilities_f64_minus_runtime_max_abs": 5.630387622801436e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:20Z"} {"id": "c-socrates-01-dessert-02", "uid": "c095", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "prompt_sha256_by_order": ["1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191"], "request_sha256": "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "compiled_prompt_sha256_by_order": ["3b30247295b5c3dec8716911428c2b205adc94f0d63fb7499beab302bd9b5b24", "47d94f6baeffe2d38319a82c77c681fea06cfc8fccfd89082f20513cf2c5b6ef", "679e4cc68c56d8be4d869df90871cd1c3221655bdd6e47c48bf51899a6f41c3a", "36007f153866a07fe2d63f094e5a85a563b8b67238b89be7a2bf35504006f7e5", "b1aee04d9d360303f87dece0913ca09699b7b9c0711e1c22e65b5bca4644f986", "46362ab9da0c35611866d89b74624466b224976f9884333b34bd20536e18016d"], "compiled_tokens_by_order": [302, 302, 302, 302, 302, 302], "choice": "hypatia", "probs": {"darwin": 0.027227298468745963, "einstein": 0.010016363345290333, "hypatia": 0.3758612615979947, "ibn_sina": 0.3316963991617687, "sagan": 0.027227298468745963, "socrates": 0.22797137895745442}, "confidence": 0.2670692775392225, "seconds": 0.17554316483377383, "seconds_all_orders": 1.053258989002643, "seconds_by_order": [0.1753875539980072, 0.17550526400009403, 0.17542194400084554, 0.17577087100289646, 0.17515493600149057, 0.17601841999930912], "argmax_stable": false, "spread": {"darwin": 0.025127616504761575, "einstein": 0.02706259493260839, "hypatia": 0.3977281708138847, "ibn_sina": 0.5466760953409349, "sagan": 0.025652604139229176, "socrates": 0.6584405064199007}, "spread_2dp": {"darwin": 0.03, "einstein": 0.03, "hypatia": 0.4, "ibn_sina": 0.5499999999999999, "sagan": 0.019999999999999997, "socrates": 0.66}, "choices_by_order": ["hypatia", "ibn_sina", "hypatia", "ibn_sina", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.027227298468745963, "einstein": 0.010016363345290333, "hypatia": 0.3758612615979947, "ibn_sina": 0.3316963991617687, "sagan": 0.027227298468745963, "socrates": 0.22797137895745442}, {"ibn_sina": 0.5770074421076441, "hypatia": 0.053670052498393996, "sagan": 0.028727508968979572, "socrates": 0.3088498277286261, "darwin": 0.015376727488528108, "einstein": 0.016368441207828103}, {"hypatia": 0.4477361073971398, "darwin": 0.010529743984324732, "sagan": 0.028622811730915868, "ibn_sina": 0.036752417759564025, "einstein": 0.028622811730915868, "socrates": 0.4477361073971398}, {"ibn_sina": 0.35935020771092163, "einstein": 0.026031241736786958, "darwin": 0.02949726130544875, "sagan": 0.033424776017977206, "hypatia": 0.1923463055179438, "socrates": 0.35935020771092163}, {"einstein": 0.03707895827789872, "sagan": 0.007772171878748031, "socrates": 0.7447507855640697, "ibn_sina": 0.12941828089528526, "darwin": 0.019846936175824104, "hypatia": 0.061132867208174214}, {"socrates": 0.8864118853773552, "ibn_sina": 0.03033134676670929, "hypatia": 0.05000793658325517, "einstein": 0.016235199999253668, "sagan": 0.01264398647273963, "darwin": 0.004369644800687174}], "prompt_tokens": 302, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.027227298468745963, "einstein": 0.010016363345290333, "hypatia": 0.3758612615979947, "ibn_sina": 0.3316963991617687, "sagan": 0.027227298468745963, "socrates": 0.22797137895745442}, "confidence": 0.2670692775392225, "choice": "hypatia"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 175.36240400295355, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 302, "compiled_prompt_sha256": "3b30247295b5c3dec8716911428c2b205adc94f0d63fb7499beab302bd9b5b24", "compiled_ids_sha256": "0da5ad6c304119d9f5f455fa27621e3de8086264734e81f84f183030922025e2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 15.625, 19.25, 19.125, 16.625, 18.75], "runtime_probabilities": [0.02722729742527008, 0.010016363114118576, 0.3758612275123596, 0.33169639110565186, 0.02722729742527008, 0.2279713749885559], "decide_seconds": 0.17162323400043533, "probabilities_f64_minus_runtime_max_abs": 3.408563509221807e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.5770074421076441, "hypatia": 0.053670052498393996, "sagan": 0.028727508968979572, "socrates": 0.3088498277286261, "darwin": 0.015376727488528108, "einstein": 0.016368441207828103}, "confidence": 0.4024681404615267, "choice": "ibn_sina"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 175.4831039979763, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 302, "compiled_prompt_sha256": "47d94f6baeffe2d38319a82c77c681fea06cfc8fccfd89082f20513cf2c5b6ef", "compiled_ids_sha256": "bc469acfacc2210432ab12540c44a63f4f17495da4ce3a8ba0122b320841a6cd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.5, 17.125, 16.5, 18.875, 15.875, 15.9375], "runtime_probabilities": [0.5770074129104614, 0.05367004871368408, 0.02872750721871853, 0.30884984135627747, 0.015376727096736431, 0.01636844128370285], "decide_seconds": 0.17185796199919423, "probabilities_f64_minus_runtime_max_abs": 2.919718267158089e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.4477361073971398, "darwin": 0.010529743984324732, "sagan": 0.028622811730915868, "ibn_sina": 0.036752417759564025, "einstein": 0.028622811730915868, "socrates": 0.4477361073971398}, "confidence": 0.3903510299615125, "choice": "hypatia"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 175.4000549990451, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 302, "compiled_prompt_sha256": "679e4cc68c56d8be4d869df90871cd1c3221655bdd6e47c48bf51899a6f41c3a", "compiled_ids_sha256": "be184a868b536337afe9e9b09ac9a7e7954a242d364f2ce1cac676ffa9a7541d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.625, 15.875, 16.875, 17.125, 16.875, 19.625], "runtime_probabilities": [0.44773608446121216, 0.010529744438827038, 0.028622809797525406, 0.03675241768360138, 0.028622809797525406, 0.44773608446121216], "decide_seconds": 0.1717686820011295, "probabilities_f64_minus_runtime_max_abs": 2.2935927657119493e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.35935020771092163, "einstein": 0.026031241736786958, "darwin": 0.02949726130544875, "sagan": 0.033424776017977206, "hypatia": 0.1923463055179438, "socrates": 0.35935020771092163}, "confidence": 0.23810459645475435, "choice": "ibn_sina"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 175.74883199995384, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 302, "compiled_prompt_sha256": "36007f153866a07fe2d63f094e5a85a563b8b67238b89be7a2bf35504006f7e5", "compiled_ids_sha256": "314d191f7ccedde6050290f5578f4ecb71047fbf321ae845a84c026df4b90fbd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.125, 16.5, 16.625, 16.75, 18.5, 19.125], "runtime_probabilities": [0.35935020446777344, 0.026031242683529854, 0.029497262090444565, 0.033424776047468185, 0.19234631955623627, 0.35935020446777344], "decide_seconds": 0.17212149999977555, "probabilities_f64_minus_runtime_max_abs": 1.403829247292876e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.03707895827789872, "sagan": 0.007772171878748031, "socrates": 0.7447507855640697, "ibn_sina": 0.12941828089528526, "darwin": 0.019846936175824104, "hypatia": 0.061132867208174214}, "confidence": 0.50179565942155, "choice": "socrates"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 175.12698599966825, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 302, "compiled_prompt_sha256": "b1aee04d9d360303f87dece0913ca09699b7b9c0711e1c22e65b5bca4644f986", "compiled_ids_sha256": "c6a725bf78557eba707736aeda915f5ff3cc72245ebae27367c3f2a38f0817bc", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 15.4375, 20.0, 18.25, 16.375, 17.5], "runtime_probabilities": [0.03707895800471306, 0.007772172335535288, 0.7447507977485657, 0.1294182687997818, 0.01984693668782711, 0.061132870614528656], "decide_seconds": 0.1714158060021873, "probabilities_f64_minus_runtime_max_abs": 1.2184495967737519e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.8864118853773552, "ibn_sina": 0.03033134676670929, "hypatia": 0.05000793658325517, "einstein": 0.016235199999253668, "sagan": 0.01264398647273963, "darwin": 0.004369644800687174}, "confidence": 0.7161415040664694, "choice": "socrates"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 175.9953000000678, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 302, "compiled_prompt_sha256": "46362ab9da0c35611866d89b74624466b224976f9884333b34bd20536e18016d", "compiled_ids_sha256": "9e74852718ba36a208eb2d3c877352b48ff0e0ecb913eff96b21ccc5d7664118", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.375, 17.0, 17.5, 16.375, 16.125, 15.0625], "runtime_probabilities": [0.8864119052886963, 0.03033134713768959, 0.05000794306397438, 0.016235200688242912, 0.012643986381590366, 0.0043696449138224125], "decide_seconds": 0.1721139299988863, "probabilities_f64_minus_runtime_max_abs": 1.9911341131439997e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:21Z"} {"id": "c-socrates-01-main-01", "uid": "c096", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "prompt_sha256_by_order": ["d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79"], "request_sha256": "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "compiled_prompt_sha256_by_order": ["1370e6b93e0997fcf22f6ff3e6473a4bccf01f168816a3e2da10413c14e6b44c", "8ac4d63f23757aa1e7bbb97c8d49528f90a81b241b6099a9916007075b7c9f98", "18b2c5ba46c76521f3fa6d8c28a5cbe620cab72d2828a21c063943ed1890e141", "38e30e3c151e24732f60c596fac287a3385502ec8dc67fdbd09c5c1d4df919ac", "d703416e934056107402f89bd2087c2fd95404a6eb696db67485e9b3c16edf8c", "370f936f1ef7b1e357129cfe9f3030c4305fefb65d4e8856c5652becf441ae71"], "compiled_tokens_by_order": [323, 323, 323, 323, 323, 323], "choice": "socrates", "probs": {"darwin": 0.0334874369538518, "einstein": 0.023015556392570954, "hypatia": 0.026080042122714083, "ibn_sina": 0.03794623738143005, "sagan": 0.015818345154023042, "socrates": 0.8636523819954102}, "confidence": 0.6584441052572991, "seconds": 0.1842145408330301, "seconds_all_orders": 1.1052872449981805, "seconds_by_order": [0.18432675499934703, 0.18419370599804097, 0.18399953700281912, 0.1843406749976566, 0.18413680699813995, 0.18428976500217686], "argmax_stable": true, "spread": {"darwin": 0.07062499037668515, "einstein": 0.03996456561228389, "hypatia": 0.03573632512513653, "ibn_sina": 0.03201053578409791, "sagan": 0.019825727714905855, "socrates": 0.18784869970122797}, "spread_2dp": {"darwin": 0.07, "einstein": 0.04, "hypatia": 0.03, "ibn_sina": 0.03, "sagan": 0.02, "socrates": 0.18999999999999995}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0334874369538518, "einstein": 0.023015556392570954, "hypatia": 0.026080042122714083, "ibn_sina": 0.03794623738143005, "sagan": 0.015818345154023042, "socrates": 0.8636523819954102}, {"ibn_sina": 0.011232260048496392, "hypatia": 0.007252092148766578, "sagan": 0.0032180963834752873, "socrates": 0.9498366281915285, "darwin": 0.0197132303063027, "einstein": 0.008747692921430647}, {"hypatia": 0.03757707075399791, "darwin": 0.029265052128743155, "sagan": 0.00950097138096719, "ibn_sina": 0.025826317857592557, "einstein": 0.04258039959567542, "socrates": 0.8552501882830237}, {"ibn_sina": 0.04298841727390311, "einstein": 0.04871225853371454, "darwin": 0.080312936788379, "sagan": 0.023010041639799816, "hypatia": 0.04298841727390311, "socrates": 0.7619879284903005}, {"einstein": 0.009214117858042829, "sagan": 0.0031843139248939616, "socrates": 0.9398655730311936, "ibn_sina": 0.022103519685359976, "darwin": 0.015191512103293114, "hypatia": 0.01044096339721665}, {"socrates": 0.9283255456998256, "ibn_sina": 0.010977881489805198, "hypatia": 0.017002875466225104, "einstein": 0.017002875466225104, "sagan": 0.017002875466225104, "darwin": 0.00968794641169385}], "prompt_tokens": 323, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0334874369538518, "einstein": 0.023015556392570954, "hypatia": 0.026080042122714083, "ibn_sina": 0.03794623738143005, "sagan": 0.015818345154023042, "socrates": 0.8636523819954102}, "confidence": 0.6584441052572991, "choice": "socrates"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 184.30115499722888, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "1370e6b93e0997fcf22f6ff3e6473a4bccf01f168816a3e2da10413c14e6b44c", "compiled_ids_sha256": "a31f962774edf9fb5117efe523b6eb2f11ca98ce116e6d7c52e3734ea7a26aa9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 16.5, 16.625, 17.0, 16.125, 20.125], "runtime_probabilities": [0.033487435430288315, 0.02301555499434471, 0.026080042123794556, 0.03794623911380768, 0.015818346291780472, 0.8636523485183716], "decide_seconds": 0.18033321599796182, "probabilities_f64_minus_runtime_max_abs": 3.3477038585694174e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.011232260048496392, "hypatia": 0.007252092148766578, "sagan": 0.0032180963834752873, "socrates": 0.9498366281915285, "darwin": 0.0197132303063027, "einstein": 0.008747692921430647}, "confidence": 0.8479936126718088, "choice": "socrates"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 184.17083600070328, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "8ac4d63f23757aa1e7bbb97c8d49528f90a81b241b6099a9916007075b7c9f98", "compiled_ids_sha256": "034dbc5dacc23f547fde305251460e794274f26f1bb333d000f71eac1d884f6b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 15.5, 14.6875, 20.375, 16.5, 15.6875], "runtime_probabilities": [0.011232259683310986, 0.007252092007547617, 0.0032180966809391975, 0.9498366713523865, 0.019713230431079865, 0.00874769315123558], "decide_seconds": 0.18027116600205773, "probabilities_f64_minus_runtime_max_abs": 4.3160857998714164e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.03757707075399791, "darwin": 0.029265052128743155, "sagan": 0.00950097138096719, "ibn_sina": 0.025826317857592557, "einstein": 0.04258039959567542, "socrates": 0.8552501882830237}, "confidence": 0.6464666014622107, "choice": "socrates"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 183.97708699922077, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "18b2c5ba46c76521f3fa6d8c28a5cbe620cab72d2828a21c063943ed1890e141", "compiled_ids_sha256": "d1c9fe1dfb59f3aaaf5898ff6ffad171a178715b36299ccdd6a18344b5d63269", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 16.875, 15.75, 16.75, 17.25, 20.25], "runtime_probabilities": [0.037577074021101, 0.02926505170762539, 0.009500971995294094, 0.025826318189501762, 0.04258039966225624, 0.8552502393722534], "decide_seconds": 0.18018922799819848, "probabilities_f64_minus_runtime_max_abs": 5.1089229735623576e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.04298841727390311, "einstein": 0.04871225853371454, "darwin": 0.080312936788379, "sagan": 0.023010041639799816, "hypatia": 0.04298841727390311, "socrates": 0.7619879284903005}, "confidence": 0.4897719561114424, "choice": "socrates"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 184.3187249978655, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "38e30e3c151e24732f60c596fac287a3385502ec8dc67fdbd09c5c1d4df919ac", "compiled_ids_sha256": "a160c69a210a975c030748f373b8c350d4f1bdb9fe6b77b7bf562a76269a02c2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.0, 17.125, 17.625, 16.375, 17.0, 19.875], "runtime_probabilities": [0.04298841953277588, 0.04871225729584694, 0.08031293749809265, 0.023010043427348137, 0.04298841953277588, 0.7619879245758057], "decide_seconds": 0.1805244839997613, "probabilities_f64_minus_runtime_max_abs": 3.914494839385441e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.009214117858042829, "sagan": 0.0031843139248939616, "socrates": 0.9398655730311936, "ibn_sina": 0.022103519685359976, "darwin": 0.015191512103293114, "hypatia": 0.01044096339721665}, "confidence": 0.8240377523613585, "choice": "socrates"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 184.11415700029465, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "d703416e934056107402f89bd2087c2fd95404a6eb696db67485e9b3c16edf8c", "compiled_ids_sha256": "85633e66aba1a86d4e40aa43553b5b4e69a92fbd21c9aef9c0906a955509e903", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.875, 14.8125, 20.5, 16.75, 16.375, 16.0], "runtime_probabilities": [0.009214119054377079, 0.0031843141186982393, 0.9398656487464905, 0.02210352197289467, 0.015191513113677502, 0.010440964251756668], "decide_seconds": 0.180202866999025, "probabilities_f64_minus_runtime_max_abs": 7.571529692107504e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.9283255456998256, "ibn_sina": 0.010977881489805198, "hypatia": 0.017002875466225104, "einstein": 0.017002875466225104, "sagan": 0.017002875466225104, "darwin": 0.00968794641169385}, "confidence": 0.7927607356101691, "choice": "socrates"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 184.2678259999957, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 323, "compiled_prompt_sha256": "370f936f1ef7b1e357129cfe9f3030c4305fefb65d4e8856c5652becf441ae71", "compiled_ids_sha256": "3828151f5a960fdb5cfad4afb2580cf6b214834f845ce0ea9c9fc791a4760a20", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.375, 15.9375, 16.375, 16.375, 16.375, 15.8125], "runtime_probabilities": [0.9283254742622375, 0.010977880097925663, 0.017002874985337257, 0.017002874985337257, 0.017002874985337257, 0.009687946178019047], "decide_seconds": 0.18033025600016117, "probabilities_f64_minus_runtime_max_abs": 7.143758806726908e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:22Z"} {"id": "c-socrates-01-main-02", "uid": "c097", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "prompt_sha256_by_order": ["b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d"], "request_sha256": "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "compiled_prompt_sha256_by_order": ["7a7cb526fa35845664c998ea567eaa951e2590b39bd2a7c881e769dec476a428", "6b3394ef3d36b8a88c04b2cd2d19c899848b44b4390af0213d9524d139bfd7d9", "c95bdf299bf010e460bb20a55fe1c0478779868b109f9c5024421795c18a8ba0", "300d217cf43c51333e1e6f2180bc908f71205ee809d827928911f8a4767deb6b", "8b631b1e7ddb293360b4bb46c71d2ee03870dda007d5250683d2791eaf0e7a95", "52541a1b16e42ec70681506c471705cfc38ab8ff7d555574f41a36b30d55f23f"], "compiled_tokens_by_order": [395, 395, 395, 395, 395, 395], "choice": "socrates", "probs": {"darwin": 0.01515332919161846, "einstein": 0.0033811647693077038, "hypatia": 0.04667553989322464, "ibn_sina": 0.09881211872936212, "sagan": 0.008634106710877922, "socrates": 0.8273437407056091}, "confidence": 0.6359412085673166, "seconds": 0.21197278949997175, "seconds_all_orders": 1.2718367369998305, "seconds_by_order": [0.2122869869999704, 0.2118408409987751, 0.21187144000214175, 0.21228366699870094, 0.21150494299945422, 0.21204885900078807], "argmax_stable": true, "spread": {"darwin": 0.038405990530495135, "einstein": 0.003998799984011843, "hypatia": 0.038341216674336806, "ibn_sina": 0.08823953731324, "sagan": 0.007433433611830738, "socrates": 0.13597304556786893}, "spread_2dp": {"darwin": 0.04, "einstein": 0.01, "hypatia": 0.04, "ibn_sina": 0.09000000000000001, "sagan": 0.01, "socrates": 0.13}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.01515332919161846, "einstein": 0.0033811647693077038, "hypatia": 0.04667553989322464, "ibn_sina": 0.09881211872936212, "sagan": 0.008634106710877922, "socrates": 0.8273437407056091}, {"ibn_sina": 0.013444667279681438, "hypatia": 0.010470717405550398, "sagan": 0.0028181553582790034, "socrates": 0.9425439435060894, "darwin": 0.025117945294286122, "einstein": 0.005604571156113562}, {"hypatia": 0.01940028351682366, "darwin": 0.013333606866618119, "sagan": 0.00382014232673406, "ibn_sina": 0.02198340125614658, "einstein": 0.006704558580785763, "socrates": 0.9347580074528918}, {"ibn_sina": 0.033751039958764294, "einstein": 0.00586505130918802, "darwin": 0.0491074733744379, "sagan": 0.007074602303548714, "hypatia": 0.033751039958764294, "socrates": 0.8704507930952968}, {"einstein": 0.0027057585967739193, "sagan": 0.0012006730990471838, "socrates": 0.963316786273478, "ibn_sina": 0.013740975967870335, "darwin": 0.010701482843942772, "hypatia": 0.008334323218887834}, {"socrates": 0.9517134495205221, "ibn_sina": 0.010572581416122124, "hypatia": 0.015383015190809635, "einstein": 0.005659095031926856, "sagan": 0.004691554564017407, "darwin": 0.01198030427660186}], "prompt_tokens": 395, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.01515332919161846, "einstein": 0.0033811647693077038, "hypatia": 0.04667553989322464, "ibn_sina": 0.09881211872936212, "sagan": 0.008634106710877922, "socrates": 0.8273437407056091}, "confidence": 0.6359412085673166, "choice": "socrates"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 212.2590670005593, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "7a7cb526fa35845664c998ea567eaa951e2590b39bd2a7c881e769dec476a428", "compiled_ids_sha256": "fba0ef685096385bfc59e52dd6d585816ee5d2066e5bbea9c00a20c6d1da8968", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 14.875, 17.5, 18.25, 15.8125, 20.375], "runtime_probabilities": [0.015153329819440842, 0.003381164977326989, 0.04667554423213005, 0.09881212562322617, 0.008634107187390327, 0.827343761920929], "decide_seconds": 0.2077247930028534, "probabilities_f64_minus_runtime_max_abs": 2.1215319834233526e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.013444667279681438, "hypatia": 0.010470717405550398, "sagan": 0.0028181553582790034, "socrates": 0.9425439435060894, "darwin": 0.025117945294286122, "einstein": 0.005604571156113562}, "confidence": 0.8327971401112385, "choice": "socrates"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 211.81626100224094, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "6b3394ef3d36b8a88c04b2cd2d19c899848b44b4390af0213d9524d139bfd7d9", "compiled_ids_sha256": "20a28d91a25d65cc9539fac61a4984b1761b0bf02bb9bd37e903487ffc0140f2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 16.0, 14.6875, 20.5, 16.875, 15.375], "runtime_probabilities": [0.013444666750729084, 0.010470718145370483, 0.002818155335262418, 0.9425439238548279, 0.025117944926023483, 0.005604570731520653], "decide_seconds": 0.20731823599999188, "probabilities_f64_minus_runtime_max_abs": 1.9651261506936635e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.01940028351682366, "darwin": 0.013333606866618119, "sagan": 0.00382014232673406, "ibn_sina": 0.02198340125614658, "einstein": 0.006704558580785763, "socrates": 0.9347580074528918}, "confidence": 0.8125508318334062, "choice": "socrates"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 211.84454999820446, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "c95bdf299bf010e460bb20a55fe1c0478779868b109f9c5024421795c18a8ba0", "compiled_ids_sha256": "6315a386fe3572227e2adc7b826a6a1623c6ea83be4e824cd61b299325b2b407", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 16.5, 15.25, 17.0, 15.8125, 20.75], "runtime_probabilities": [0.019400283694267273, 0.013333606533706188, 0.0038201422430574894, 0.021983401849865913, 0.006704558152705431, 0.9347580075263977], "decide_seconds": 0.2074177150025207, "probabilities_f64_minus_runtime_max_abs": 5.937193341132296e-10}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.033751039958764294, "einstein": 0.00586505130918802, "darwin": 0.0491074733744379, "sagan": 0.007074602303548714, "hypatia": 0.033751039958764294, "socrates": 0.8704507930952968}, "confidence": 0.6859614339847877, "choice": "socrates"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 212.2589969985711, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "300d217cf43c51333e1e6f2180bc908f71205ee809d827928911f8a4767deb6b", "compiled_ids_sha256": "2a36b86eacc3196d5219396270e81598da67bf34fd583de0dcb05fd710af56ac", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.25, 15.5, 17.625, 15.6875, 17.25, 20.5], "runtime_probabilities": [0.03375103697180748, 0.005865050945430994, 0.049107473343610764, 0.007074602413922548, 0.03375103697180748, 0.8704507350921631], "decide_seconds": 0.20772749300158466, "probabilities_f64_minus_runtime_max_abs": 5.800313374049182e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.0027057585967739193, "sagan": 0.0012006730990471838, "socrates": 0.963316786273478, "ibn_sina": 0.013740975967870335, "darwin": 0.010701482843942772, "hypatia": 0.008334323218887834}, "confidence": 0.8842240511689653, "choice": "socrates"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 211.47990300232777, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "8b631b1e7ddb293360b4bb46c71d2ee03870dda007d5250683d2791eaf0e7a95", "compiled_ids_sha256": "72191dec5f384150c83643ceaae301013348ed8050ea27e46a363d8d127ee808", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0, 14.1875, 20.875, 16.625, 16.375, 16.125], "runtime_probabilities": [0.0027057589031755924, 0.001200673170387745, 0.963316798210144, 0.01374097540974617, 0.01070148404687643, 0.008334323763847351], "decide_seconds": 0.20675666000170168, "probabilities_f64_minus_runtime_max_abs": 1.1936665988976358e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.9517134495205221, "ibn_sina": 0.010572581416122124, "hypatia": 0.015383015190809635, "einstein": 0.005659095031926856, "sagan": 0.004691554564017407, "darwin": 0.01198030427660186}, "confidence": 0.8510606901257739, "choice": "socrates"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 212.02447899850085, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 395, "compiled_prompt_sha256": "52541a1b16e42ec70681506c471705cfc38ab8ff7d555574f41a36b30d55f23f", "compiled_ids_sha256": "5e60fa068aef24d30a8d982a4dfeeb7ef924cb25b06ba0a127552ec31504ea4a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.75, 16.25, 16.625, 15.625, 15.4375, 16.375], "runtime_probabilities": [0.9517134428024292, 0.010572582483291626, 0.015383015386760235, 0.005659095011651516, 0.004691554699093103, 0.01198030449450016], "decide_seconds": 0.20753411400073674, "probabilities_f64_minus_runtime_max_abs": 6.71809285979208e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:23Z"} {"id": "c-socrates-01-main-02", "uid": "c098", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "prompt_sha256_by_order": ["7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53"], "request_sha256": "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "compiled_prompt_sha256_by_order": ["ae667f183d359d40c186509ad74566d963a401022b071a698cabca10ad469305", "49640ee5153760731e8f22301aa321ba4a52e2aee38de52860b386825a269840", "44a75d78e0aaaecb479ef43e098bc75854d559cb678c48f8e2cee707acf08adc", "0144f56f396253c962fe0b30ed9f396a34ce328a66dd5ae3c9ff4ced79fe878b", "e4cf79a383e4270bd75dbad0aced9f5f9b58e7bfd06ab3e39727d6af5a53f92e", "1897c93ff7a27e62ef324efcaada8d4989f2f3b41b3a912beb3f9588c66e9713"], "compiled_tokens_by_order": [297, 297, 297, 297, 297, 297], "choice": "socrates", "probs": {"darwin": 0.06312901033248118, "einstein": 0.03828968028396546, "hypatia": 0.1336441159226044, "ibn_sina": 0.3205955678708105, "sagan": 0.08105925379724889, "socrates": 0.36328237179288964}, "confidence": 0.16031706233630905, "seconds": 0.17524294216673297, "seconds_all_orders": 1.0514576530003978, "seconds_by_order": [0.17501387700031046, 0.17541004400118254, 0.17501936799817486, 0.17524532500101486, 0.1753349350001372, 0.17543410399957793], "argmax_stable": true, "spread": {"darwin": 0.0554822061875776, "einstein": 0.026791839084866625, "hypatia": 0.3190237729348019, "ibn_sina": 0.29035187079256974, "sagan": 0.06372528676756607, "socrates": 0.4947649880612139}, "spread_2dp": {"darwin": 0.049999999999999996, "einstein": 0.03, "hypatia": 0.32, "ibn_sina": 0.29000000000000004, "sagan": 0.06, "socrates": 0.5}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.06312901033248118, "einstein": 0.03828968028396546, "hypatia": 0.1336441159226044, "ibn_sina": 0.3205955678708105, "sagan": 0.08105925379724889, "socrates": 0.36328237179288964}, {"ibn_sina": 0.033270106660443374, "hypatia": 0.04840771955437695, "sagan": 0.025910785120022463, "socrates": 0.8580473598541035, "darwin": 0.022866187611954848, "einstein": 0.011497841199098838}, {"hypatia": 0.10265119446761203, "darwin": 0.033325963561192024, "sagan": 0.029410059618399053, "ibn_sina": 0.042791384248180075, "einstein": 0.033325963561192024, "socrates": 0.7584954345434248}, {"ibn_sina": 0.07153557155258594, "einstein": 0.023224199808436798, "darwin": 0.026316466086645042, "sagan": 0.04916563138027677, "hypatia": 0.3632876084082103, "socrates": 0.46647052276384526}, {"einstein": 0.02684737333109785, "sagan": 0.017333967029682814, "socrates": 0.7845952570087624, "ibn_sina": 0.08269573168364021, "darwin": 0.04426383547340838, "hypatia": 0.04426383547340838}, {"socrates": 0.7799952280004476, "ibn_sina": 0.03024369707824075, "hypatia": 0.10556087510464689, "einstein": 0.02668996899425424, "sagan": 0.049863426677506845, "darwin": 0.007646804144903583}], "prompt_tokens": 297, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.06312901033248118, "einstein": 0.03828968028396546, "hypatia": 0.1336441159226044, "ibn_sina": 0.3205955678708105, "sagan": 0.08105925379724889, "socrates": 0.36328237179288964}, "confidence": 0.16031706233630905, "choice": "socrates"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 174.98784700001124, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 297, "compiled_prompt_sha256": "ae667f183d359d40c186509ad74566d963a401022b071a698cabca10ad469305", "compiled_ids_sha256": "455a8afb9852c21fd5125d01820fb2e27419fa8e5daf8123220881908a4db405", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 16.625, 17.875, 18.75, 17.375, 18.875], "runtime_probabilities": [0.0631290003657341, 0.03828967735171318, 0.13364411890506744, 0.32059553265571594, 0.08105925470590591, 0.36328235268592834], "decide_seconds": 0.17124970700024278, "probabilities_f64_minus_runtime_max_abs": 3.521509456261285e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.033270106660443374, "hypatia": 0.04840771955437695, "sagan": 0.025910785120022463, "socrates": 0.8580473598541035, "darwin": 0.022866187611954848, "einstein": 0.011497841199098838}, "confidence": 0.6519855748992792, "choice": "socrates"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 175.38327500005835, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 297, "compiled_prompt_sha256": "49640ee5153760731e8f22301aa321ba4a52e2aee38de52860b386825a269840", "compiled_ids_sha256": "70f658f6e03afeae0583917abcec524f436afde8d7cd7e3c0acb04b1fc085bdb", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.75, 17.125, 16.5, 20.0, 16.375, 15.6875], "runtime_probabilities": [0.033270109444856644, 0.04840772598981857, 0.025910787284374237, 0.8580474257469177, 0.02286618947982788, 0.011497842147946358], "decide_seconds": 0.1717424930029665, "probabilities_f64_minus_runtime_max_abs": 6.589281420499304e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.10265119446761203, "darwin": 0.033325963561192024, "sagan": 0.029410059618399053, "ibn_sina": 0.042791384248180075, "einstein": 0.033325963561192024, "socrates": 0.7584954345434248}, "confidence": 0.49289170809933547, "choice": "socrates"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 174.99733799922979, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 297, "compiled_prompt_sha256": "44a75d78e0aaaecb479ef43e098bc75854d559cb678c48f8e2cee707acf08adc", "compiled_ids_sha256": "87f83b83d25d1cd65863c60d394606cbe0c6ce75d748516fb7f0bb833675156f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.0, 16.875, 16.75, 17.125, 16.875, 20.0], "runtime_probabilities": [0.1026511862874031, 0.03332596272230148, 0.02941005863249302, 0.04279138520359993, 0.03332596272230148, 0.7584953904151917], "decide_seconds": 0.17132757600120385, "probabilities_f64_minus_runtime_max_abs": 4.41282331786752e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.07153557155258594, "einstein": 0.023224199808436798, "darwin": 0.026316466086645042, "sagan": 0.04916563138027677, "hypatia": 0.3632876084082103, "socrates": 0.46647052276384526}, "confidence": 0.3060080446549467, "choice": "socrates"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 175.22330599967972, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 297, "compiled_prompt_sha256": "0144f56f396253c962fe0b30ed9f396a34ce328a66dd5ae3c9ff4ced79fe878b", "compiled_ids_sha256": "6d5872db9660d96f58948862d9dd482b37cccff50ea6b2c6d2b069366fa683a9", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.25, 16.125, 16.25, 16.875, 18.875, 19.125], "runtime_probabilities": [0.07153557240962982, 0.023224197328090668, 0.026316465809941292, 0.04916562885046005, 0.3632875978946686, 0.46647050976753235], "decide_seconds": 0.17165523400035454, "probabilities_f64_minus_runtime_max_abs": 1.2996312914737018e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.02684737333109785, "sagan": 0.017333967029682814, "socrates": 0.7845952570087624, "ibn_sina": 0.08269573168364021, "darwin": 0.04426383547340838, "hypatia": 0.04426383547340838}, "confidence": 0.5312622094038271, "choice": "socrates"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 175.31224500271492, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 297, "compiled_prompt_sha256": "e4cf79a383e4270bd75dbad0aced9f5f9b58e7bfd06ab3e39727d6af5a53f92e", "compiled_ids_sha256": "7d46106b95b4889d5ddfa6c63eb0979a5cbc2e457659e063755bcdfe321748e4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 15.9375, 19.75, 17.5, 16.875, 16.875], "runtime_probabilities": [0.026847373694181442, 0.017333967611193657, 0.7845953106880188, 0.08269573003053665, 0.04426383972167969, 0.04426383972167969], "decide_seconds": 0.1715656240012322, "probabilities_f64_minus_runtime_max_abs": 5.367925637322912e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.7799952280004476, "ibn_sina": 0.03024369707824075, "hypatia": 0.10556087510464689, "einstein": 0.02668996899425424, "sagan": 0.049863426677506845, "darwin": 0.007646804144903583}, "confidence": 0.5420977006353755, "choice": "socrates"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 175.41276500196545, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 297, "compiled_prompt_sha256": "1897c93ff7a27e62ef324efcaada8d4989f2f3b41b3a912beb3f9588c66e9713", "compiled_ids_sha256": "cdb2a8f6114e92846e6ae309493852e0a849d99637df8360978b8f3da0b230dd", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.0, 16.75, 18.0, 16.625, 17.25, 15.375], "runtime_probabilities": [0.7799952030181885, 0.03024369478225708, 0.10556086897850037, 0.026689967140555382, 0.04986342415213585, 0.0076468042097985744], "decide_seconds": 0.17183007200219436, "probabilities_f64_minus_runtime_max_abs": 2.498225915026353e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:24Z"} {"id": "c-socrates-01-main-02", "uid": "c099", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "prompt_sha256_by_order": ["d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755"], "request_sha256": "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "compiled_prompt_sha256_by_order": ["241005cdbb0ac4de9754451ae5ce2a6d308deccb51feb8a49be378b2807af5d6", "eaec06d2423a7ba9b0c3f7fff3dca302237d8a5f27b823252a18121d4e1a7ba3", "a689d1686bffde1b551f8fd49c12f38a50a6ff9ad72e67027af2c3be66b3a033", "f6b1cf8376e48fa3129bc369507669dded1555c084c8a6ba61dcc3bdd8391cd4", "8067d15b1b4637576aecaa56150d4addfd6b218c11de084a93d142545a8513c0", "df18e9943e030e2e856eb69266f0acd5e414c0fd09095675a3c1f595e96c8d9d"], "compiled_tokens_by_order": [368, 368, 368, 368, 368, 368], "choice": "socrates", "probs": {"darwin": 0.020271133755179496, "einstein": 0.0031086790449346615, "hypatia": 0.09084891859394896, "ibn_sina": 0.11665232055322602, "sagan": 0.008450265758357311, "socrates": 0.7606686822943536}, "confidence": 0.5457307495984214, "seconds": 0.19434731183294693, "seconds_all_orders": 1.1660838709976815, "seconds_by_order": [0.19442298600188224, 0.19455712600029074, 0.19422544799817842, 0.19419255799948587, 0.19425466699976823, 0.19443108599807601], "argmax_stable": true, "spread": {"darwin": 0.018590149677354013, "einstein": 0.005186000555774043, "hypatia": 0.17397242099596877, "ibn_sina": 0.09485298734530177, "sagan": 0.0072377296657332325, "socrates": 0.1987822241758863}, "spread_2dp": {"darwin": 0.02, "einstein": 0.01, "hypatia": 0.16999999999999998, "ibn_sina": 0.09999999999999999, "sagan": 0.01, "socrates": 0.20000000000000007}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.020271133755179496, "einstein": 0.0031086790449346615, "hypatia": 0.09084891859394896, "ibn_sina": 0.11665232055322602, "sagan": 0.008450265758357311, "socrates": 0.7606686822943536}, {"ibn_sina": 0.05830839972384555, "hypatia": 0.010132480556297487, "sagan": 0.005773313392898337, "socrates": 0.9120968326362152, "darwin": 0.010785969407271626, "einstein": 0.0029030042834718374}, {"hypatia": 0.18410490155226625, "darwin": 0.00522265405472719, "sagan": 0.006706020548837003, "ibn_sina": 0.06772840829997112, "einstein": 0.00808900483924588, "socrates": 0.7281490107049526}, {"ibn_sina": 0.06918609720401686, "einstein": 0.006045413799948689, "darwin": 0.009967202322159493, "sagan": 0.0050118240716070855, "hypatia": 0.16596881926872215, "socrates": 0.7438206433335457}, {"einstein": 0.004232166811558751, "sagan": 0.0012125360926240783, "socrates": 0.9138935608852156, "ibn_sina": 0.045500081196742856, "darwin": 0.010807216554760217, "hypatia": 0.024354438459098626}, {"socrates": 0.926931234880839, "ibn_sina": 0.021799333207924255, "hypatia": 0.0407265161404225, "einstein": 0.004569388472681995, "sagan": 0.004292543220306783, "darwin": 0.0016809840778254826}], "prompt_tokens": 368, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.020271133755179496, "einstein": 0.0031086790449346615, "hypatia": 0.09084891859394896, "ibn_sina": 0.11665232055322602, "sagan": 0.008450265758357311, "socrates": 0.7606686822943536}, "confidence": 0.5457307495984214, "choice": "socrates"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 194.39729599980637, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 368, "compiled_prompt_sha256": "241005cdbb0ac4de9754451ae5ce2a6d308deccb51feb8a49be378b2807af5d6", "compiled_ids_sha256": "7479577962a5718990d77f717d29e7cc80e2c6bc96a00d11d0d6eae9329107e2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 14.5, 17.875, 18.125, 15.5, 20.0], "runtime_probabilities": [0.02027113363146782, 0.0031086790841072798, 0.09084892272949219, 0.11665232479572296, 0.008450266905128956, 0.7606686949729919], "decide_seconds": 0.19007901000077254, "probabilities_f64_minus_runtime_max_abs": 1.267863836140748e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.05830839972384555, "hypatia": 0.010132480556297487, "sagan": 0.005773313392898337, "socrates": 0.9120968326362152, "darwin": 0.010785969407271626, "einstein": 0.0029030042834718374}, "confidence": 0.7813680910826405, "choice": "socrates"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 194.53255600092234, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 368, "compiled_prompt_sha256": "eaec06d2423a7ba9b0c3f7fff3dca302237d8a5f27b823252a18121d4e1a7ba3", "compiled_ids_sha256": "8d305155a1c25196f54bb00326f4df85877094db1e198ad74a90f597c034047f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 15.75, 15.1875, 20.25, 15.8125, 14.5], "runtime_probabilities": [0.058308400213718414, 0.010132482275366783, 0.00577331380918622, 0.9120969176292419, 0.010785969905555248, 0.0029030046425759792], "decide_seconds": 0.19027517900030944, "probabilities_f64_minus_runtime_max_abs": 8.499302672770881e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.18410490155226625, "darwin": 0.00522265405472719, "sagan": 0.006706020548837003, "ibn_sina": 0.06772840829997112, "einstein": 0.00808900483924588, "socrates": 0.7281490107049526}, "confidence": 0.5396309394737165, "choice": "socrates"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 194.20200799868326, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 368, "compiled_prompt_sha256": "a689d1686bffde1b551f8fd49c12f38a50a6ff9ad72e67027af2c3be66b3a033", "compiled_ids_sha256": "3681413d94547e23758d234b431d9df1ca4568d2e26ae808ceda1258c6f3521f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.625, 15.0625, 15.3125, 17.625, 15.5, 20.0], "runtime_probabilities": [0.18410490453243256, 0.005222653970122337, 0.006706020329147577, 0.06772840768098831, 0.008089005015790462, 0.7281489968299866], "decide_seconds": 0.18999201100086793, "probabilities_f64_minus_runtime_max_abs": 1.387496606497507e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.06918609720401686, "einstein": 0.006045413799948689, "darwin": 0.009967202322159493, "sagan": 0.0050118240716070855, "hypatia": 0.16596881926872215, "socrates": 0.7438206433335457}, "confidence": 0.5499609172462454, "choice": "socrates"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 194.16870799977914, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 368, "compiled_prompt_sha256": "f6b1cf8376e48fa3129bc369507669dded1555c084c8a6ba61dcc3bdd8391cd4", "compiled_ids_sha256": "51dfe3e0726a0d00ea433504f874c4f7735dcadf85dea872d59e972a8b78c724", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.75, 15.3125, 15.8125, 15.125, 18.625, 20.125], "runtime_probabilities": [0.06918609142303467, 0.006045414134860039, 0.009967202320694923, 0.005011823959648609, 0.16596882045269012, 0.7438206076622009], "decide_seconds": 0.18984657100008917, "probabilities_f64_minus_runtime_max_abs": 3.5671344766896596e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.004232166811558751, "sagan": 0.0012125360926240783, "socrates": 0.9138935608852156, "ibn_sina": 0.045500081196742856, "darwin": 0.010807216554760217, "hypatia": 0.024354438459098626}, "confidence": 0.780347624815647, "choice": "socrates"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 194.23006699798862, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 368, "compiled_prompt_sha256": "8067d15b1b4637576aecaa56150d4addfd6b218c11de084a93d142545a8513c0", "compiled_ids_sha256": "ba3e8f90ccb2f79a64382c745df6f4d594567d0e08baaa8be0129ceb828f9d22", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0, 13.75, 20.375, 17.375, 15.9375, 16.75], "runtime_probabilities": [0.004232167266309261, 0.0012125361245125532, 0.9138935804367065, 0.04550008103251457, 0.010807216167449951, 0.024354437366127968], "decide_seconds": 0.18982067199976882, "probabilities_f64_minus_runtime_max_abs": 1.9551490981761788e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.926931234880839, "ibn_sina": 0.021799333207924255, "hypatia": 0.0407265161404225, "einstein": 0.004569388472681995, "sagan": 0.004292543220306783, "darwin": 0.0016809840778254826}, "confidence": 0.8086505064627132, "choice": "socrates"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 194.4077459993423, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 368, "compiled_prompt_sha256": "df18e9943e030e2e856eb69266f0acd5e414c0fd09095675a3c1f595e96c8d9d", "compiled_ids_sha256": "6b1036d66a025f372cfece7382b11c77c1ac584a2277666034702f986e836522", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.375, 16.625, 17.25, 15.0625, 15.0, 14.0625], "runtime_probabilities": [0.9269312024116516, 0.02179933339357376, 0.040726516395807266, 0.004569388460367918, 0.004292543511837721, 0.0016809840453788638], "decide_seconds": 0.19020337899928563, "probabilities_f64_minus_runtime_max_abs": 3.246918733879056e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:25Z"} {"id": "c-socrates-01-main-02", "uid": "c100", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "prompt_sha256_by_order": ["b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2"], "request_sha256": "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "compiled_prompt_sha256_by_order": ["5a8c5648b9712c46350b863549c1ceed5801044f634662c32f1ac342b250aa4c", "2715c49d2ec43081a62bb93258e0ddb4d67c44598ed54685cf64c50a9a589424", "3a9ec945decbf700df228adc3d5522a7edff7ccf4ee4e66eb8181b8759d75a71", "80caaf9373c0b7ea823fa5b533d765c439fe9f6ac78199808a18c3b81c553d2e", "f5e6e09cffc1d28b199c0e08bd0888ac630076a3722af6e42e376e39a05786e7", "27e95771a2772083c0cfe4b58b1285d9da010e45929422d7a7cdbe0438f82348"], "compiled_tokens_by_order": [298, 298, 298, 298, 298, 298], "choice": "sagan", "probs": {"darwin": 0.11313472832957773, "einstein": 0.14526786668524047, "hypatia": 0.11313472832957773, "ibn_sina": 0.06861968141014833, "sagan": 0.34847949639852943, "socrates": 0.21136349884692632}, "confidence": 0.07742798820300378, "seconds": 0.17546497249956397, "seconds_all_orders": 1.0527898349973839, "seconds_by_order": [0.1757566520027467, 0.17566421299852664, 0.17534934500145027, 0.17488048899758724, 0.17561979299716768, 0.17551934299990535], "argmax_stable": false, "spread": {"darwin": 0.11815339452030679, "einstein": 0.29705585272681323, "hypatia": 0.10339647218233164, "ibn_sina": 0.168556277011864, "sagan": 0.271898944608615, "socrates": 0.21687082758975693}, "spread_2dp": {"darwin": 0.12000000000000001, "einstein": 0.29000000000000004, "hypatia": 0.10999999999999999, "ibn_sina": 0.17, "sagan": 0.26999999999999996, "socrates": 0.21999999999999997}, "choices_by_order": ["sagan", "einstein", "einstein", "sagan", "socrates", "einstein"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.11313472832957773, "einstein": 0.14526786668524047, "hypatia": 0.11313472832957773, "ibn_sina": 0.06861968141014833, "sagan": 0.34847949639852943, "socrates": 0.21136349884692632}, {"ibn_sina": 0.09258100950819041, "hypatia": 0.1347045739911031, "sagan": 0.1959939986668573, "socrates": 0.1347045739911031, "darwin": 0.11887636931112595, "einstein": 0.3231394745316202}, {"hypatia": 0.06783253932516094, "darwin": 0.04114251482687301, "sagan": 0.07686433700390077, "ibn_sina": 0.06783253932516094, "einstein": 0.4423237194120537, "socrates": 0.30400435010685073}, {"ibn_sina": 0.08818116715444545, "einstein": 0.16474410905794779, "darwin": 0.08818116715444545, "sagan": 0.3487632816125158, "hypatia": 0.1453861659626977, "socrates": 0.16474410905794779}, {"einstein": 0.18818475168230883, "sagan": 0.10072803901096407, "socrates": 0.35157540158086004, "ibn_sina": 0.18818475168230883, "darwin": 0.12933736226319223, "hypatia": 0.041989693780366065}, {"socrates": 0.18622978254383518, "ibn_sina": 0.01962847467044485, "hypatia": 0.12799373293395208, "einstein": 0.3479230383910019, "sagan": 0.3070410037178805, "darwin": 0.011183967742885437}], "prompt_tokens": 298, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.11313472832957773, "einstein": 0.14526786668524047, "hypatia": 0.11313472832957773, "ibn_sina": 0.06861968141014833, "sagan": 0.34847949639852943, "socrates": 0.21136349884692632}, "confidence": 0.07742798820300378, "choice": "sagan"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 175.7303820013476, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 298, "compiled_prompt_sha256": "5a8c5648b9712c46350b863549c1ceed5801044f634662c32f1ac342b250aa4c", "compiled_ids_sha256": "b7bb323dfd3806fa1b596523b49ddfd4bc24913a34b61cf4015eef0afa9175a5", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.0, 18.25, 18.0, 17.5, 19.125, 18.625], "runtime_probabilities": [0.1131347268819809, 0.14526785910129547, 0.1131347268819809, 0.06861967593431473, 0.3484794795513153, 0.21136349439620972], "decide_seconds": 0.17197987100007595, "probabilities_f64_minus_runtime_max_abs": 1.684721412376078e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.09258100950819041, "hypatia": 0.1347045739911031, "sagan": 0.1959939986668573, "socrates": 0.1347045739911031, "darwin": 0.11887636931112595, "einstein": 0.3231394745316202}, "confidence": 0.05232599383843961, "choice": "einstein"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 175.64114299966604, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 298, "compiled_prompt_sha256": "2715c49d2ec43081a62bb93258e0ddb4d67c44598ed54685cf64c50a9a589424", "compiled_ids_sha256": "cffb8432c20632ab08338fbc36efe71f8f4f981a448d75e3144a7838f7d8db85", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.875, 18.25, 18.625, 18.25, 18.125, 19.125], "runtime_probabilities": [0.09258101880550385, 0.1347045749425888, 0.19599400460720062, 0.1347045749425888, 0.1188763827085495, 0.323139488697052], "decide_seconds": 0.17200257100193994, "probabilities_f64_minus_runtime_max_abs": 1.4165431827528607e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.06783253932516094, "darwin": 0.04114251482687301, "sagan": 0.07686433700390077, "ibn_sina": 0.06783253932516094, "einstein": 0.4423237194120537, "socrates": 0.30400435010685073}, "confidence": 0.20954067291378642, "choice": "einstein"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 175.32747500081314, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 298, "compiled_prompt_sha256": "3a9ec945decbf700df228adc3d5522a7edff7ccf4ee4e66eb8181b8759d75a71", "compiled_ids_sha256": "3cc932c6fa6a6ebf003196cf9406c5ed21a22ecce851734ffd2fe25380151d32", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 17.125, 17.75, 17.625, 19.5, 19.125], "runtime_probabilities": [0.06783254444599152, 0.04114251583814621, 0.0768643394112587, 0.06783254444599152, 0.4423237442970276, 0.30400437116622925], "decide_seconds": 0.17169107399968198, "probabilities_f64_minus_runtime_max_abs": 2.4884973914840458e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.08818116715444545, "einstein": 0.16474410905794779, "darwin": 0.08818116715444545, "sagan": 0.3487632816125158, "hypatia": 0.1453861659626977, "socrates": 0.16474410905794779}, "confidence": 0.06784969731816337, "choice": "sagan"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 174.85841899906518, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 298, "compiled_prompt_sha256": "80caaf9373c0b7ea823fa5b533d765c439fe9f6ac78199808a18c3b81c553d2e", "compiled_ids_sha256": "ce647e2cd3d97f071c8cf32eaa7ddb768691ca7251af97c552b35a416866fe9b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.5, 18.125, 17.5, 18.875, 18.0, 18.125], "runtime_probabilities": [0.08818116784095764, 0.16474412381649017, 0.08818116784095764, 0.34876328706741333, 0.14538615942001343, 0.16474412381649017], "decide_seconds": 0.17129529600060778, "probabilities_f64_minus_runtime_max_abs": 1.4758542388193518e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.18818475168230883, "sagan": 0.10072803901096407, "socrates": 0.35157540158086004, "ibn_sina": 0.18818475168230883, "darwin": 0.12933736226319223, "hypatia": 0.041989693780366065}, "confidence": 0.09304928768936715, "choice": "socrates"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 175.59700299898395, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 298, "compiled_prompt_sha256": "f5e6e09cffc1d28b199c0e08bd0888ac630076a3722af6e42e376e39a05786e7", "compiled_ids_sha256": "04afb2fece8b1c7c6f4d7c3408d35bcd36749267ddc32787bc5a29fea7b54184", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.5, 17.875, 19.125, 18.5, 18.125, 17.0], "runtime_probabilities": [0.18818475306034088, 0.10072804242372513, 0.35157540440559387, 0.18818475306034088, 0.1293373703956604, 0.04198969528079033], "decide_seconds": 0.1718345320005028, "probabilities_f64_minus_runtime_max_abs": 8.132468171107021e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.18622978254383518, "ibn_sina": 0.01962847467044485, "hypatia": 0.12799373293395208, "einstein": 0.3479230383910019, "sagan": 0.3070410037178805, "darwin": 0.011183967742885437}, "confidence": 0.19999406805729736, "choice": "einstein"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 175.4974729992682, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 298, "compiled_prompt_sha256": "27e95771a2772083c0cfe4b58b1285d9da010e45929422d7a7cdbe0438f82348", "compiled_ids_sha256": "e6b41d31c07ac2c04e603e631bcabb0dde14772c1b815d2a7205d21c80798ceb", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.5, 16.25, 18.125, 19.125, 19.0, 15.6875], "runtime_probabilities": [0.18622978031635284, 0.019628476351499557, 0.12799374759197235, 0.34792304039001465, 0.3070409893989563, 0.01118396781384945], "decide_seconds": 0.17187367199949222, "probabilities_f64_minus_runtime_max_abs": 1.4658020269786576e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:26Z"} {"id": "c-socrates-01-main-03", "uid": "c101", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "prompt_sha256_by_order": ["087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b"], "request_sha256": "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "compiled_prompt_sha256_by_order": ["aa8a76f665cd82962eddc0c1cfe089528415b48ad3760ba5f5715122b5f5c9b1", "752ba78f33d5a48f1c426bacf0099db95d6283dcafeef13c404c42674ae7e498", "0ea387a6e148bc3af33e032902215e3d7192a6bcce27aa972b4003b859b57a3c", "7d6e81fb87cb2bd9f961687c450633f3ea84686727cd1830a704b84e559427e0", "c362ea45c5237f201996344a7fb1da3574abbfb717ef1a14d8d12717b123b784", "f533ca7dc314f0448dd4327d2fc5474b352157a292ad23d17cc8b3beb05741a7"], "compiled_tokens_by_order": [340, 340, 340, 340, 340, 340], "choice": "sagan", "probs": {"darwin": 0.02481816994197398, "einstein": 0.021901958101610143, "hypatia": 0.34260426812640027, "ibn_sina": 0.059535694715278226, "sagan": 0.43991258813999984, "socrates": 0.11122732097473746}, "confidence": 0.26558369389751935, "seconds": 0.18613302733441137, "seconds_all_orders": 1.1167981640064681, "seconds_by_order": [0.18635863899908145, 0.18604956100170966, 0.1859877420029079, 0.18604731100276695, 0.18635900899971602, 0.18599590200028615], "argmax_stable": true, "spread": {"darwin": 0.032213439162549734, "einstein": 0.15365286152368665, "hypatia": 0.3075713824176012, "ibn_sina": 0.06964334818182319, "sagan": 0.19800261466362906, "socrates": 0.07531238568331947}, "spread_2dp": {"darwin": 0.03, "einstein": 0.16, "hypatia": 0.31, "ibn_sina": 0.06999999999999999, "sagan": 0.19999999999999996, "socrates": 0.07}, "choices_by_order": ["sagan", "sagan", "sagan", "sagan", "sagan", "sagan"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.02481816994197398, "einstein": 0.021901958101610143, "hypatia": 0.34260426812640027, "ibn_sina": 0.059535694715278226, "sagan": 0.43991258813999984, "socrates": 0.11122732097473746}, {"ibn_sina": 0.07846869213553326, "hypatia": 0.06111147888174228, "sagan": 0.5798095681991735, "socrates": 0.11417127337351965, "darwin": 0.03706598560215781, "einstein": 0.12937300180787353}, {"hypatia": 0.19892967228683076, "darwin": 0.010542886390017886, "sagan": 0.47720747608584957, "ibn_sina": 0.07318213667329745, "einstein": 0.1755548196252968, "socrates": 0.06458300893870754}, {"ibn_sina": 0.07259795983752924, "einstein": 0.030263332162411675, "darwin": 0.016198794404997174, "sagan": 0.47339816460551826, "hypatia": 0.36868286129934347, "socrates": 0.03885888769020018}, {"einstein": 0.083000113410253, "sagan": 0.4215093560307345, "socrates": 0.10657425520672989, "ibn_sina": 0.09405145011519932, "darwin": 0.03920647744871447, "hypatia": 0.2556583477883687}, {"socrates": 0.1093895236625092, "ibn_sina": 0.02440810193337613, "hypatia": 0.3369434538803785, "einstein": 0.14045892870202695, "sagan": 0.38180695353554445, "darwin": 0.006993038286164741}], "prompt_tokens": 340, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.02481816994197398, "einstein": 0.021901958101610143, "hypatia": 0.34260426812640027, "ibn_sina": 0.059535694715278226, "sagan": 0.43991258813999984, "socrates": 0.11122732097473746}, "confidence": 0.26558369389751935, "choice": "sagan"}}, "usage": {"input_tokens": 340, "output_tokens": 1}, "latency_ms": 186.33265899916296, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 340, "compiled_prompt_sha256": "aa8a76f665cd82962eddc0c1cfe089528415b48ad3760ba5f5715122b5f5c9b1", "compiled_ids_sha256": "86a40936a666d5fd48b556179b1723278e9807945ccc3cb11a9396580f0dd5a4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 16.25, 19.0, 17.25, 19.25, 17.875], "runtime_probabilities": [0.024818168953061104, 0.02190195582807064, 0.34260424971580505, 0.059535689651966095, 0.4399125576019287, 0.11122731119394302], "decide_seconds": 0.18220909099909477, "probabilities_f64_minus_runtime_max_abs": 3.053807112740614e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.07846869213553326, "hypatia": 0.06111147888174228, "sagan": 0.5798095681991735, "socrates": 0.11417127337351965, "darwin": 0.03706598560215781, "einstein": 0.12937300180787353}, "confidence": 0.2627285766660924, "choice": "sagan"}}, "usage": {"input_tokens": 340, "output_tokens": 1}, "latency_ms": 186.0263220005436, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 340, "compiled_prompt_sha256": "752ba78f33d5a48f1c426bacf0099db95d6283dcafeef13c404c42674ae7e498", "compiled_ids_sha256": "5459472ab449027ca2803ae13899f655ebd87eb091b0d2dd09405639570009c3", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.25, 17.0, 19.25, 17.625, 16.5, 17.75], "runtime_probabilities": [0.0784686952829361, 0.06111147999763489, 0.5798096060752869, 0.11417126655578613, 0.037065986543893814, 0.12937301397323608], "decide_seconds": 0.18200704300033976, "probabilities_f64_minus_runtime_max_abs": 3.787611335681618e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.19892967228683076, "darwin": 0.010542886390017886, "sagan": 0.47720747608584957, "ibn_sina": 0.07318213667329745, "einstein": 0.1755548196252968, "socrates": 0.06458300893870754}, "confidence": 0.22087785349144118, "choice": "sagan"}}, "usage": {"input_tokens": 340, "output_tokens": 1}, "latency_ms": 185.96404200070538, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 340, "compiled_prompt_sha256": "0ea387a6e148bc3af33e032902215e3d7192a6bcce27aa972b4003b859b57a3c", "compiled_ids_sha256": "9fab0fceceb800bddcf101132b2871a0786b128ec068cd02eec060322221956c", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.375, 15.4375, 19.25, 17.375, 18.25, 17.25], "runtime_probabilities": [0.19892965257167816, 0.010542885400354862, 0.4772074222564697, 0.0731821358203888, 0.17555482685565948, 0.06458300352096558], "decide_seconds": 0.18188481299876003, "probabilities_f64_minus_runtime_max_abs": 5.382937984244052e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.07259795983752924, "einstein": 0.030263332162411675, "darwin": 0.016198794404997174, "sagan": 0.47339816460551826, "hypatia": 0.36868286129934347, "socrates": 0.03885888769020018}, "confidence": 0.3240428285339635, "choice": "sagan"}}, "usage": {"input_tokens": 340, "output_tokens": 1}, "latency_ms": 186.02363099853392, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 340, "compiled_prompt_sha256": "7d6e81fb87cb2bd9f961687c450633f3ea84686727cd1830a704b84e559427e0", "compiled_ids_sha256": "221af17b634b5509eba3321d659ff40244dbf72a25ae37b973e755200ab33acf", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.25, 16.375, 15.75, 19.125, 18.875, 16.625], "runtime_probabilities": [0.07259796559810638, 0.030263330787420273, 0.016198793426156044, 0.4733981788158417, 0.368682861328125, 0.03885888680815697], "decide_seconds": 0.18183549500099616, "probabilities_f64_minus_runtime_max_abs": 1.421032341797357e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.083000113410253, "sagan": 0.4215093560307345, "socrates": 0.10657425520672989, "ibn_sina": 0.09405145011519932, "darwin": 0.03920647744871447, "hypatia": 0.2556583477883687}, "confidence": 0.15873204584994327, "choice": "sagan"}}, "usage": {"input_tokens": 340, "output_tokens": 1}, "latency_ms": 186.33494900132064, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 340, "compiled_prompt_sha256": "c362ea45c5237f201996344a7fb1da3574abbfb717ef1a14d8d12717b123b784", "compiled_ids_sha256": "a01d19e0e7fbf06b82ca23cb2e27420bcad2a1a648449f1f290a7c6d25067685", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.375, 19.0, 17.625, 17.5, 16.625, 18.5], "runtime_probabilities": [0.08300010859966278, 0.42150935530662537, 0.10657425969839096, 0.09405145794153214, 0.039206478744745255, 0.255658358335495], "decide_seconds": 0.18219443199996022, "probabilities_f64_minus_runtime_max_abs": 1.0547126283455555e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.1093895236625092, "ibn_sina": 0.02440810193337613, "hypatia": 0.3369434538803785, "einstein": 0.14045892870202695, "sagan": 0.38180695353554445, "darwin": 0.006993038286164741}, "confidence": 0.23134305833004654, "choice": "sagan"}}, "usage": {"input_tokens": 340, "output_tokens": 1}, "latency_ms": 185.97298199892975, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 340, "compiled_prompt_sha256": "f533ca7dc314f0448dd4327d2fc5474b352157a292ad23d17cc8b3beb05741a7", "compiled_ids_sha256": "53702ffe5108955515bdfb2ddaecf0b23e58a9d927d626522e8e7901770cd8d8", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 16.125, 18.75, 17.875, 18.875, 14.875], "runtime_probabilities": [0.109389528632164, 0.02440810017287731, 0.33694344758987427, 0.14045894145965576, 0.3818069398403168, 0.006993038579821587], "decide_seconds": 0.18195336300050258, "probabilities_f64_minus_runtime_max_abs": 1.3695227674137556e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:27Z"} {"id": "c-socrates-01-main-03", "uid": "c102", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "prompt_sha256_by_order": ["da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb"], "request_sha256": "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "compiled_prompt_sha256_by_order": ["302cd7d856db5d188e298657e794f2abd5a5c6239133269be866a306600f9f17", "878c56ee00af1549d43d0014299394f15d4be3e597a4917593617077f919ecb1", "b403e3a6bbc5f872856896adacda3ce13eb1b0014257ef2496f6e08ef82a92a5", "fdd319adf77f2aea4125d2a2e8fc438e60c0b15948b8a31576a379579411b3c5", "cb5d2bf455d73adc8c5ac8b875b4d226401ca3180f9c99dfffb61606b9034de0", "d02043f78697ca10320140700161350c84ec358c36ac6c5f0abd5a2ade19c74b"], "compiled_tokens_by_order": [380, 380, 380, 380, 380, 380], "choice": "ibn_sina", "probs": {"darwin": 0.011326369511739752, "einstein": 0.0034543475331349186, "hypatia": 0.13798342817346546, "ibn_sina": 0.5457350449176567, "sagan": 0.009389890128502979, "socrates": 0.29211091973550024}, "confidence": 0.3986681869390365, "seconds": 0.19608418166717456, "seconds_all_orders": 1.1765050900030474, "seconds_by_order": [0.1961820729993633, 0.19646196000030614, 0.19556279700191226, 0.1959014650019526, 0.19638991100146086, 0.19600688399805222], "argmax_stable": false, "spread": {"darwin": 0.02499168335121342, "einstein": 0.013410013267816132, "hypatia": 0.536150706673458, "ibn_sina": 0.5035905210462268, "sagan": 0.006830458202800227, "socrates": 0.5101596982139455}, "spread_2dp": {"darwin": 0.019999999999999997, "einstein": 0.02, "hypatia": 0.54, "ibn_sina": 0.51, "sagan": 0.01, "socrates": 0.51}, "choices_by_order": ["ibn_sina", "hypatia", "socrates", "hypatia", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.011326369511739752, "einstein": 0.0034543475331349186, "hypatia": 0.13798342817346546, "ibn_sina": 0.5457350449176567, "sagan": 0.009389890128502979, "socrates": 0.29211091973550024}, {"ibn_sina": 0.0421445238714299, "hypatia": 0.6592512728551588, "sagan": 0.005033445582152495, "socrates": 0.2748168170780289, "darwin": 0.015504103890258141, "einstein": 0.0032498367229717414}, {"hypatia": 0.1282574467625153, "darwin": 0.01052801234100584, "sagan": 0.0049730808967357884, "ibn_sina": 0.395060748517107, "einstein": 0.01351823543305371, "socrates": 0.44766247604958226}, {"ibn_sina": 0.3268345073388576, "einstein": 0.014360106054786449, "darwin": 0.030400344756542682, "sagan": 0.01050607802301307, "hypatia": 0.4196638144737093, "socrates": 0.1982351493530908}, {"einstein": 0.010635758604085218, "sagan": 0.0036756198202128423, "socrates": 0.5124595479175963, "ibn_sina": 0.24206875006905382, "darwin": 0.0175354014405872, "hypatia": 0.21362492214846457}, {"socrates": 0.7083948475670363, "ibn_sina": 0.13949121614044482, "hypatia": 0.12310056618170089, "einstein": 0.016659849990787873, "sagan": 0.006944858714700812, "darwin": 0.005408661405329263}], "prompt_tokens": 380, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.011326369511739752, "einstein": 0.0034543475331349186, "hypatia": 0.13798342817346546, "ibn_sina": 0.5457350449176567, "sagan": 0.009389890128502979, "socrates": 0.29211091973550024}, "confidence": 0.3986681869390365, "choice": "ibn_sina"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 196.15581299876794, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 380, "compiled_prompt_sha256": "302cd7d856db5d188e298657e794f2abd5a5c6239133269be866a306600f9f17", "compiled_ids_sha256": "064b264cbea6255d0964374c8390710e5279938b074c351063f0f26916c2824f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.875, 14.6875, 18.375, 19.75, 15.6875, 19.125], "runtime_probabilities": [0.011326368898153305, 0.003454347373917699, 0.13798342645168304, 0.5457350611686707, 0.009389889426529408, 0.2921109199523926], "decide_seconds": 0.19168280699886964, "probabilities_f64_minus_runtime_max_abs": 1.625101397895179e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0421445238714299, "hypatia": 0.6592512728551588, "sagan": 0.005033445582152495, "socrates": 0.2748168170780289, "darwin": 0.015504103890258141, "einstein": 0.0032498367229717414}, "confidence": 0.5127940507850718, "choice": "hypatia"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 196.43799999903422, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 380, "compiled_prompt_sha256": "878c56ee00af1549d43d0014299394f15d4be3e597a4917593617077f919ecb1", "compiled_ids_sha256": "bd9a0dfc1c36995835124b983ac3cccc71f6ddb227d9fafbe99a3dcd51c84a6f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.25, 20.0, 15.125, 19.125, 16.25, 14.6875], "runtime_probabilities": [0.042144522070884705, 0.6592512726783752, 0.005033445544540882, 0.2748168110847473, 0.01550410408526659, 0.0032498366199433804], "decide_seconds": 0.19205933500052197, "probabilities_f64_minus_runtime_max_abs": 5.993281593230648e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.1282574467625153, "darwin": 0.01052801234100584, "sagan": 0.0049730808967357884, "ibn_sina": 0.395060748517107, "einstein": 0.01351823543305371, "socrates": 0.44766247604958226}, "confidence": 0.3734689907775055, "choice": "socrates"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 195.53881699903286, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 380, "compiled_prompt_sha256": "b403e3a6bbc5f872856896adacda3ce13eb1b0014257ef2496f6e08ef82a92a5", "compiled_ids_sha256": "0a148a20c59656a2bb852819d80b63ec2e6a27360847d4a5408ab470e318ce97", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [18.5, 16.0, 15.25, 19.625, 16.25, 19.75], "runtime_probabilities": [0.12825743854045868, 0.010528012178838253, 0.004973080940544605, 0.3950607180595398, 0.013518234714865685, 0.44766244292259216], "decide_seconds": 0.1911948319975636, "probabilities_f64_minus_runtime_max_abs": 3.312699009550002e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.3268345073388576, "einstein": 0.014360106054786449, "darwin": 0.030400344756542682, "sagan": 0.01050607802301307, "hypatia": 0.4196638144737093, "socrates": 0.1982351493530908}, "confidence": 0.2936028712692996, "choice": "hypatia"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 195.87802499881946, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 380, "compiled_prompt_sha256": "fdd319adf77f2aea4125d2a2e8fc438e60c0b15948b8a31576a379579411b3c5", "compiled_ids_sha256": "eb004f6cadad59ebf9a174777f1c5594f626f7a6913fd626368744ce692b45ab", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [19.25, 16.125, 16.875, 15.8125, 19.5, 18.75], "runtime_probabilities": [0.326834499835968, 0.014360105618834496, 0.030400346964597702, 0.010506078600883484, 0.41966381669044495, 0.1982351541519165], "decide_seconds": 0.1914967990014702, "probabilities_f64_minus_runtime_max_abs": 7.502889587396311e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.010635758604085218, "sagan": 0.0036756198202128423, "socrates": 0.5124595479175963, "ibn_sina": 0.24206875006905382, "darwin": 0.0175354014405872, "hypatia": 0.21362492214846457}, "confidence": 0.3550748252934127, "choice": "socrates"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 196.36503099900438, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 380, "compiled_prompt_sha256": "cb5d2bf455d73adc8c5ac8b875b4d226401ca3180f9c99dfffb61606b9034de0", "compiled_ids_sha256": "01dd6d124fff090dfc66e3f02e6dd3d270b911b630deeeabbe1f0f04569ca867", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.75, 14.6875, 19.625, 18.875, 16.25, 18.75], "runtime_probabilities": [0.010635758750140667, 0.003675619838759303, 0.5124595761299133, 0.24206876754760742, 0.01753540150821209, 0.21362492442131042], "decide_seconds": 0.19183511600203929, "probabilities_f64_minus_runtime_max_abs": 2.8212317038089907e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.7083948475670363, "ibn_sina": 0.13949121614044482, "hypatia": 0.12310056618170089, "einstein": 0.016659849990787873, "sagan": 0.006944858714700812, "darwin": 0.005408661405329263}, "confidence": 0.49333908236711754, "choice": "socrates"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 195.9832850006933, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 380, "compiled_prompt_sha256": "d02043f78697ca10320140700161350c84ec358c36ac6c5f0abd5a2ade19c74b", "compiled_ids_sha256": "9e231a171471ce8e90baadb5861bf4c2f421c5985f1eef02045a1f6bc829141f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.0, 18.375, 18.25, 16.25, 15.375, 15.125], "runtime_probabilities": [0.7083948850631714, 0.1394912153482437, 0.12310055643320084, 0.016659850254654884, 0.006944859400391579, 0.005408661440014839], "decide_seconds": 0.19160226800158853, "probabilities_f64_minus_runtime_max_abs": 3.749613508574612e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:29Z"} {"id": "c-socrates-01-starter-01", "uid": "c103", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "prompt_sha256_by_order": ["ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c"], "request_sha256": "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "compiled_prompt_sha256_by_order": ["7ba348fbd49cae74d977fc56e0a7c6c7e8609ef11e84d162768dc1a823298c56", "3f5a56ceaad52ea0d53e09f9207d50a24b2c91cc0f25122104a03b375067966d", "a5c584d3428cacb1c6d335557737106622f032c16affb109e5eacf7e649ad131", "02d48ce06fd008951db9a3bc5e3a5bba5372afe6ad859080d27399c8b623e642", "116ea19dea447cdf35718566bd027424c4d502f6007db60def9a1e508f969933", "3c991d8fb09f07f5f4233d8c042b256cdf894c97b925bfc524776aecbcc8e0c7"], "compiled_tokens_by_order": [267, 267, 267, 267, 267, 267], "choice": "socrates", "probs": {"darwin": 0.016314984724209124, "einstein": 0.005296700045628277, "hypatia": 0.03453882293218625, "ibn_sina": 0.044348726507404575, "sagan": 0.00873278202974568, "socrates": 0.8907679837608262}, "confidence": 0.724425464862452, "seconds": 0.17218158449941257, "seconds_all_orders": 1.0330895069964754, "seconds_by_order": [0.17248368699802086, 0.17182982200029073, 0.17187145199932274, 0.17186759199830703, 0.17227423900112626, 0.17276271499940776], "argmax_stable": true, "spread": {"darwin": 0.013292861792092322, "einstein": 0.008004112535130942, "hypatia": 0.07204802842996118, "ibn_sina": 0.03849284494158792, "sagan": 0.007259419992507321, "socrates": 0.13050858373682261}, "spread_2dp": {"darwin": 0.02, "einstein": 0.01, "hypatia": 0.06999999999999999, "ibn_sina": 0.04, "sagan": 0.01, "socrates": 0.13}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.016314984724209124, "einstein": 0.005296700045628277, "hypatia": 0.03453882293218625, "ibn_sina": 0.044348726507404575, "sagan": 0.00873278202974568, "socrates": 0.8907679837608262}, {"ibn_sina": 0.01941038802713388, "hypatia": 0.02824194793388772, "sagan": 0.004907714663193821, "socrates": 0.9352448700245485, "darwin": 0.010389632023511584, "einstein": 0.001805447327724636}, {"hypatia": 0.05851965105827814, "darwin": 0.007439938728221162, "sagan": 0.004512550945049553, "ibn_sina": 0.010825046974847251, "einstein": 0.003301452797816191, "socrates": 0.9154013594957877}, {"ibn_sina": 0.04679879273822267, "einstein": 0.009809559862855578, "darwin": 0.017216313720035507, "sagan": 0.009215228675617297, "hypatia": 0.08743165534589295, "socrates": 0.829528449657376}, {"einstein": 0.0038895900114352087, "sagan": 0.0019558086831099764, "socrates": 0.9517512956136924, "ibn_sina": 0.019752967960898055, "darwin": 0.0072667108149325815, "hypatia": 0.015383626915931776}, {"socrates": 0.9600370333941987, "ibn_sina": 0.008305947796634752, "hypatia": 0.02257790696372169, "einstein": 0.0021000724836764554, "sagan": 0.0030555874338251653, "darwin": 0.003923451927943185}], "prompt_tokens": 267, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.016314984724209124, "einstein": 0.005296700045628277, "hypatia": 0.03453882293218625, "ibn_sina": 0.044348726507404575, "sagan": 0.00873278202974568, "socrates": 0.8907679837608262}, "confidence": 0.724425464862452, "choice": "socrates"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 172.4594369989063, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 267, "compiled_prompt_sha256": "7ba348fbd49cae74d977fc56e0a7c6c7e8609ef11e84d162768dc1a823298c56", "compiled_ids_sha256": "279db28cb15e54380a75c8b3d192304b55a16372b39c8bf6c63f78d6061d4052", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 15.0, 16.875, 17.125, 15.5, 20.125], "runtime_probabilities": [0.01631498709321022, 0.00529670063406229, 0.03453882783651352, 0.04434873163700104, 0.008732783608138561, 0.8907681107521057], "decide_seconds": 0.1689878149991273, "probabilities_f64_minus_runtime_max_abs": 1.2699127949211686e-07}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.01941038802713388, "hypatia": 0.02824194793388772, "sagan": 0.004907714663193821, "socrates": 0.9352448700245485, "darwin": 0.010389632023511584, "einstein": 0.001805447327724636}, "confidence": 0.818719052559767, "choice": "socrates"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 171.80790199927287, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 267, "compiled_prompt_sha256": "3f5a56ceaad52ea0d53e09f9207d50a24b2c91cc0f25122104a03b375067966d", "compiled_ids_sha256": "601c34aa442f08593a2e3803f078d24eeef568aa02631b4b4060bbc617bc38ee", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.375, 16.75, 15.0, 20.25, 15.75, 14.0], "runtime_probabilities": [0.01941038854420185, 0.028241949155926704, 0.004907714668661356, 0.9352449178695679, 0.010389632545411587, 0.0018054473912343383], "decide_seconds": 0.1685418569977628, "probabilities_f64_minus_runtime_max_abs": 4.784501939081309e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.05851965105827814, "darwin": 0.007439938728221162, "sagan": 0.004512550945049553, "ibn_sina": 0.010825046974847251, "einstein": 0.003301452797816191, "socrates": 0.9154013594957877}, "confidence": 0.7903144018761266, "choice": "socrates"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 171.8506620018161, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 267, "compiled_prompt_sha256": "a5c584d3428cacb1c6d335557737106622f032c16affb109e5eacf7e649ad131", "compiled_ids_sha256": "b2f8d83a099b8fa9ecac061d465635552b171fa65cde774a53500a42450b8c32", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.625, 15.5625, 15.0625, 15.9375, 14.75, 20.375], "runtime_probabilities": [0.058519650250673294, 0.007439939305186272, 0.00451255077496171, 0.010825046338140965, 0.0033014528453350067, 0.9154013395309448], "decide_seconds": 0.16859740700238035, "probabilities_f64_minus_runtime_max_abs": 1.9964842890018986e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.04679879273822267, "einstein": 0.009809559862855578, "darwin": 0.017216313720035507, "sagan": 0.009215228675617297, "hypatia": 0.08743165534589295, "socrates": 0.829528449657376}, "confidence": 0.6261343643858754, "choice": "socrates"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 171.846353001456, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 267, "compiled_prompt_sha256": "02d48ce06fd008951db9a3bc5e3a5bba5372afe6ad859080d27399c8b623e642", "compiled_ids_sha256": "6d65abb3870c492f6868afb827346392da2b104c5435416ed7602e8907480b96", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [17.125, 15.5625, 16.125, 15.5, 17.75, 20.0], "runtime_probabilities": [0.04679879546165466, 0.00980955921113491, 0.017216313630342484, 0.00921522919088602, 0.0874316543340683, 0.8295284509658813], "decide_seconds": 0.1685713080005371, "probabilities_f64_minus_runtime_max_abs": 2.7234319938118645e-09}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.0038895900114352087, "sagan": 0.0019558086831099764, "socrates": 0.9517512956136924, "ibn_sina": 0.019752967960898055, "darwin": 0.0072667108149325815, "hypatia": 0.015383626915931776}, "confidence": 0.8558002537446241, "choice": "socrates"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 172.2530689985433, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 267, "compiled_prompt_sha256": "116ea19dea447cdf35718566bd027424c4d502f6007db60def9a1e508f969933", "compiled_ids_sha256": "dda3eb65361d6c9c41a6d01f4925e09bfac28339a36c3b7511c4dd62c2740d15", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.0, 14.3125, 20.5, 16.625, 15.625, 16.375], "runtime_probabilities": [0.0038895902689546347, 0.001955808838829398, 0.9517513513565063, 0.019752968102693558, 0.007266710978001356, 0.015383628197014332], "decide_seconds": 0.16885281600116286, "probabilities_f64_minus_runtime_max_abs": 5.574281392028979e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.9600370333941987, "ibn_sina": 0.008305947796634752, "hypatia": 0.02257790696372169, "einstein": 0.0021000724836764554, "sagan": 0.0030555874338251653, "darwin": 0.003923451927943185}, "confidence": 0.8789372217666344, "choice": "socrates"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 172.74153499965905, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 267, "compiled_prompt_sha256": "3c991d8fb09f07f5f4233d8c042b256cdf894c97b925bfc524776aecbcc8e0c7", "compiled_ids_sha256": "cefc78669241588f0f8cbfa7a9289d44dc5a78227a0565f397573587aae36023", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.5, 15.75, 16.75, 14.375, 14.75, 15.0], "runtime_probabilities": [0.9600370526313782, 0.008305948227643967, 0.022577907890081406, 0.002100072568282485, 0.0030555876437574625, 0.003923451993614435], "decide_seconds": 0.16948503999810782, "probabilities_f64_minus_runtime_max_abs": 1.9237179516018443e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:30Z"} {"id": "c-socrates-01-starter-02", "uid": "c104", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "prompt_sha256_by_order": ["caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d"], "request_sha256": "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "compiled_prompt_sha256_by_order": ["09a3150ed8b79aaef3b11c5a10df28e718cebf9189825e56bd6872be31be436d", "15b4cd0fceefa103657844be6d1ba89db7dfa950d506eb26c21b6c0e921ad99b", "81f07b4c59c6b0e2a3ac361cc52b529a009e086ae13d368060b101bacf0501f0", "297cf3e7b0cb93d214dff7d93af8d03747074f169f88e60e5870669375322034", "57b0a5c48171094738510bfa2a4aae666897260d2743d7f7cecd03c9a8caecee", "513bcaa62245dda4f0599b7a24d1bf790b2259b02178ab05dab7f744b736f641"], "compiled_tokens_by_order": [315, 315, 315, 315, 315, 315], "choice": "socrates", "probs": {"darwin": 0.0107564812054241, "einstein": 0.001549618041509407, "hypatia": 0.012188690038368711, "ibn_sina": 0.00395708829482254, "sagan": 0.003280541419618715, "socrates": 0.9682675810002564}, "confidence": 0.8970987892918806, "seconds": 0.17745789550038657, "seconds_all_orders": 1.0647473730023194, "seconds_by_order": [0.17735143000027165, 0.1770392719990923, 0.17755208699963987, 0.17744190899975365, 0.17774621700300486, 0.17761645800055703], "argmax_stable": true, "spread": {"darwin": 0.009188429363483052, "einstein": 0.0020878004272441826, "hypatia": 0.0212947629604483, "ibn_sina": 0.0029486882960091066, "sagan": 0.011471946080979203, "socrates": 0.0352193076756816}, "spread_2dp": {"darwin": 0.01, "einstein": 0.0, "hypatia": 0.02, "ibn_sina": 0.0, "sagan": 0.01, "socrates": 0.040000000000000036}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0107564812054241, "einstein": 0.001549618041509407, "hypatia": 0.012188690038368711, "ibn_sina": 0.00395708829482254, "sagan": 0.003280541419618715, "socrates": 0.9682675810002564}, {"ibn_sina": 0.004011463806869526, "hypatia": 0.0013863247959513348, "sagan": 0.001780076273785961, "socrates": 0.9815728301108898, "darwin": 0.008492268945783931, "einstein": 0.0027570360667194447}, {"hypatia": 0.019816848948415792, "darwin": 0.005677622282374938, "sagan": 0.012019526466108236, "ibn_sina": 0.004421736679497259, "einstein": 0.0032350116618068223, "socrates": 0.9548292539617969}, {"ibn_sina": 0.0034782574608895785, "einstein": 0.00144995343010827, "darwin": 0.006104531954841159, "sagan": 0.0018617770572725918, "hypatia": 0.022681087756399633, "socrates": 0.9644243923404888}, {"einstein": 0.0013982954989075372, "sagan": 0.0005475803851290334, "socrates": 0.9900485616374785, "ibn_sina": 0.0026123599131296785, "darwin": 0.0026123599131296785, "hypatia": 0.0027808426522256836}, {"socrates": 0.9797858073466668, "ibn_sina": 0.0014730483834881518, "hypatia": 0.010884437142033064, "einstein": 0.00114721123456264, "sagan": 0.005141444051308421, "darwin": 0.001568051841941048}], "prompt_tokens": 315, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0107564812054241, "einstein": 0.001549618041509407, "hypatia": 0.012188690038368711, "ibn_sina": 0.00395708829482254, "sagan": 0.003280541419618715, "socrates": 0.9682675810002564}, "confidence": 0.8970987892918806, "choice": "socrates"}}, "usage": {"input_tokens": 315, "output_tokens": 1}, "latency_ms": 177.3209599996335, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 315, "compiled_prompt_sha256": "09a3150ed8b79aaef3b11c5a10df28e718cebf9189825e56bd6872be31be436d", "compiled_ids_sha256": "2fae6a509382acdcb1398799443b33dcf79f533516bd7b6f1ae007a1d10494dc", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.25, 14.3125, 16.375, 15.25, 15.0625, 20.75], "runtime_probabilities": [0.010756482370197773, 0.0015496182022616267, 0.012188690714538097, 0.003957088571041822, 0.0032805416267365217, 0.9682676196098328], "decide_seconds": 0.17353648899734253, "probabilities_f64_minus_runtime_max_abs": 3.8609576313852756e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.004011463806869526, "hypatia": 0.0013863247959513348, "sagan": 0.001780076273785961, "socrates": 0.9815728301108898, "darwin": 0.008492268945783931, "einstein": 0.0027570360667194447}, "confidence": 0.9344038804052086, "choice": "socrates"}}, "usage": {"input_tokens": 315, "output_tokens": 1}, "latency_ms": 177.0167320028122, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 315, "compiled_prompt_sha256": "15b4cd0fceefa103657844be6d1ba89db7dfa950d506eb26c21b6c0e921ad99b", "compiled_ids_sha256": "5be6ca73d195d1470fed5f5b6987f664dac9260eba28bada034ea920b50c2e7b", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.375, 14.3125, 14.5625, 20.875, 16.125, 15.0], "runtime_probabilities": [0.00401146337389946, 0.0013863247586414218, 0.0017800760688260198, 0.9815727472305298, 0.00849226862192154, 0.002757035894319415], "decide_seconds": 0.17331745000046794, "probabilities_f64_minus_runtime_max_abs": 8.288036001946608e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.019816848948415792, "darwin": 0.005677622282374938, "sagan": 0.012019526466108236, "ibn_sina": 0.004421736679497259, "einstein": 0.0032350116618068223, "socrates": 0.9548292539617969}, "confidence": 0.8622235595627868, "choice": "socrates"}}, "usage": {"input_tokens": 315, "output_tokens": 1}, "latency_ms": 177.53008699946804, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 315, "compiled_prompt_sha256": "81f07b4c59c6b0e2a3ac361cc52b529a009e086ae13d368060b101bacf0501f0", "compiled_ids_sha256": "7b39dab83d1ca7724c52829cee4e2579aaa9f8c847ef78366aa994a37b94d6d4", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.875, 15.625, 16.375, 15.375, 15.0625, 20.75], "runtime_probabilities": [0.019816849380731583, 0.005677622277289629, 0.012019527144730091, 0.004421737045049667, 0.003235011827200651, 0.9548293352127075], "decide_seconds": 0.17381843700059108, "probabilities_f64_minus_runtime_max_abs": 8.12509106662418e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0034782574608895785, "einstein": 0.00144995343010827, "darwin": 0.006104531954841159, "sagan": 0.0018617770572725918, "hypatia": 0.022681087756399633, "socrates": 0.9644243923404888}, "confidence": 0.8923917424339092, "choice": "socrates"}}, "usage": {"input_tokens": 315, "output_tokens": 1}, "latency_ms": 177.4190889991587, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 315, "compiled_prompt_sha256": "297cf3e7b0cb93d214dff7d93af8d03747074f169f88e60e5870669375322034", "compiled_ids_sha256": "a6cadf0dd1e21e1c9c35a4acfcef547b396a9320c21c11ec55a31f3ccbce5f64", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.125, 14.25, 15.6875, 14.5, 17.0, 20.75], "runtime_probabilities": [0.003478257218375802, 0.0014499532990157604, 0.006104531232267618, 0.0018617769237607718, 0.022681087255477905, 0.9644243121147156], "decide_seconds": 0.17359180799758178, "probabilities_f64_minus_runtime_max_abs": 8.022577324684477e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.0013982954989075372, "sagan": 0.0005475803851290334, "socrates": 0.9900485616374785, "ibn_sina": 0.0026123599131296785, "darwin": 0.0026123599131296785, "hypatia": 0.0027808426522256836}, "confidence": 0.9605730134200575, "choice": "socrates"}}, "usage": {"input_tokens": 315, "output_tokens": 1}, "latency_ms": 177.72209699978703, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 315, "compiled_prompt_sha256": "57b0a5c48171094738510bfa2a4aae666897260d2743d7f7cecd03c9a8caecee", "compiled_ids_sha256": "20684a84b7109c4f8c28890a046cd6d8e6afc3f5d60cfb0c06b1ec120e333447", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.3125, 13.375, 20.875, 14.9375, 14.9375, 15.0], "runtime_probabilities": [0.0013982955133542418, 0.0005475803627632558, 0.9900485277175903, 0.0026123598217964172, 0.0026123598217964172, 0.0027808428276330233], "decide_seconds": 0.17379306699876906, "probabilities_f64_minus_runtime_max_abs": 3.391988812051494e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.9797858073466668, "ibn_sina": 0.0014730483834881518, "hypatia": 0.010884437142033064, "einstein": 0.00114721123456264, "sagan": 0.005141444051308421, "darwin": 0.001568051841941048}, "confidence": 0.9309021831717216, "choice": "socrates"}}, "usage": {"input_tokens": 315, "output_tokens": 1}, "latency_ms": 177.59399799979292, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 315, "compiled_prompt_sha256": "513bcaa62245dda4f0599b7a24d1bf790b2259b02178ab05dab7f744b736f641", "compiled_ids_sha256": "d9b510cab54a1f427220ab3ebe3f09ad1bd8c19efbbc6dbf8a6bd01f5fd9a6d7", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.875, 14.375, 16.375, 14.125, 15.625, 14.4375], "runtime_probabilities": [0.9797858595848083, 0.0014730483526363969, 0.010884438641369343, 0.0011472112964838743, 0.005141444504261017, 0.0015680519863963127], "decide_seconds": 0.17379723700287286, "probabilities_f64_minus_runtime_max_abs": 5.2238141590166265e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:31Z"} {"id": "c-socrates-01-starter-03", "uid": "c105", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "prompt_sha256_by_order": ["b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f"], "request_sha256": "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "compiled_prompt_sha256_by_order": ["1b43313b82313432c0e61c8d24490e5ee29de32faa2f8898725bc668a978fa6f", "ae6b89e50a5148483737380bffe9acace8cc964e5580e839cc8895265059d463", "87525d5ed757dbe3f1a1f394ce48f41cfeff3a70beab9ab2ee7aec44eca6793f", "93158928479c6a94dccd8832caae1cd3689d9b803d2128164772c4d869e1d61c", "4baed2eca5b55d49908e1c2c5bcfdc217d52427193c34234525c058ceca51e6b", "53075d3ffd6a2d6a6792062761c6c2a82284c8b143a93e98e928e589e597923a"], "compiled_tokens_by_order": [335, 335, 335, 335, 335, 335], "choice": "socrates", "probs": {"darwin": 0.0014076946377844155, "einstein": 0.0004864862492704877, "hypatia": 0.0004293226081312028, "ibn_sina": 0.0004864862492704877, "sagan": 0.0004864862492704877, "socrates": 0.9967035240062729}, "confidence": 0.9849335019403302, "seconds": 0.18554084383322333, "seconds_all_orders": 1.11324506299934, "seconds_by_order": [0.1857634039988625, 0.1854107360013586, 0.18498256000020774, 0.18573108399868943, 0.18577851399822976, 0.185578765001992], "argmax_stable": true, "spread": {"darwin": 0.001786386191188343, "einstein": 0.0016220672526214537, "hypatia": 0.0015564167255338235, "ibn_sina": 0.000504365573647566, "sagan": 0.00048548124543273, "socrates": 0.005530384887840412}, "spread_2dp": {"darwin": 0.0, "einstein": 0.0, "hypatia": 0.0, "ibn_sina": 0.0, "sagan": 0.0, "socrates": 0.010000000000000009}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.0014076946377844155, "einstein": 0.0004864862492704877, "hypatia": 0.0004293226081312028, "ibn_sina": 0.0004864862492704877, "sagan": 0.0004864862492704877, "socrates": 0.9967035240062729}, {"ibn_sina": 0.00045773679413141957, "hypatia": 0.0002450088503124866, "sagan": 0.0001792523041188635, "socrates": 0.9982852475233068, "darwin": 0.0005877456778179069, "einstein": 0.0002450088503124866}, {"hypatia": 0.00180142557584631, "darwin": 0.0019176075436413112, "sagan": 0.0005848372579609405, "ibn_sina": 0.000549403759748522, "einstein": 0.00180142557584631, "socrates": 0.9933453002869566}, {"ibn_sina": 0.0007076051801544716, "einstein": 0.0006647335495515935, "darwin": 0.000551083468409685, "sagan": 0.0006647335495515935, "hypatia": 0.0010295594620641218, "socrates": 0.9963822847902685}, {"einstein": 0.0006653600323528608, "sagan": 0.00031429382481417084, "socrates": 0.9973213322707658, "ibn_sina": 0.0005871781676549858, "darwin": 0.0006250479058662742, "hypatia": 0.00048678779854579614}, {"socrates": 0.998875685174797, "ibn_sina": 0.00020323960650690562, "hypatia": 0.00031478366044037433, "einstein": 0.0001793583232248562, "sagan": 0.00029571188257792924, "darwin": 0.00013122135245296823}], "prompt_tokens": 335, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0014076946377844155, "einstein": 0.0004864862492704877, "hypatia": 0.0004293226081312028, "ibn_sina": 0.0004864862492704877, "sagan": 0.0004864862492704877, "socrates": 0.9967035240062729}, "confidence": 0.9849335019403302, "choice": "socrates"}}, "usage": {"input_tokens": 335, "output_tokens": 1}, "latency_ms": 185.73832400215906, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 335, "compiled_prompt_sha256": "1b43313b82313432c0e61c8d24490e5ee29de32faa2f8898725bc668a978fa6f", "compiled_ids_sha256": "8635156608b5843ff3e2d0a3a4e35ad222d413b6cce00d96311035874e998319", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.1875, 13.125, 13.0, 13.125, 13.125, 20.75], "runtime_probabilities": [0.0014076948864385486, 0.0004864863585680723, 0.0004293226811569184, 0.0004864863585680723, 0.0004864863585680723, 0.9967036843299866], "decide_seconds": 0.1817313150022528, "probabilities_f64_minus_runtime_max_abs": 1.6032371363827735e-07}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.00045773679413141957, "hypatia": 0.0002450088503124866, "sagan": 0.0001792523041188635, "socrates": 0.9982852475233068, "darwin": 0.0005877456778179069, "einstein": 0.0002450088503124866}, "confidence": 0.9915023428355346, "choice": "socrates"}}, "usage": {"input_tokens": 335, "output_tokens": 1}, "latency_ms": 185.38389700188418, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 335, "compiled_prompt_sha256": "ae6b89e50a5148483737380bffe9acace8cc964e5580e839cc8895265059d463", "compiled_ids_sha256": "fe51e5ec80c56190dd90407e8d0fd3470d1968081556cf8f1521d4d83d86fc5a", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [13.3125, 12.6875, 12.375, 21.0, 13.5625, 12.6875], "runtime_probabilities": [0.00045773680903948843, 0.00024500887957401574, 0.0001792522962205112, 0.9982852339744568, 0.0005877456860616803, 0.00024500887957401574], "decide_seconds": 0.18147802699968452, "probabilities_f64_minus_runtime_max_abs": 1.3548850041900096e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.00180142557584631, "darwin": 0.0019176075436413112, "sagan": 0.0005848372579609405, "ibn_sina": 0.000549403759748522, "einstein": 0.00180142557584631, "socrates": 0.9933453002869566}, "confidence": 0.9721640978277513, "choice": "socrates"}}, "usage": {"input_tokens": 335, "output_tokens": 1}, "latency_ms": 184.95975000041653, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 335, "compiled_prompt_sha256": "87525d5ed757dbe3f1a1f394ce48f41cfeff3a70beab9ab2ee7aec44eca6793f", "compiled_ids_sha256": "1062e99eefbe42425eac67a4c9a9f9e58fc4bd82ae7f4a114703f46f0d3d2542", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.1875, 14.25, 13.0625, 13.0, 14.1875, 20.5], "runtime_probabilities": [0.0018014254746958613, 0.001917607500217855, 0.0005848372238688171, 0.0005494037177413702, 0.0018014254746958613, 0.9933452606201172], "decide_seconds": 0.18109592000109842, "probabilities_f64_minus_runtime_max_abs": 3.966683936873494e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0007076051801544716, "einstein": 0.0006647335495515935, "darwin": 0.000551083468409685, "sagan": 0.0006647335495515935, "hypatia": 0.0010295594620641218, "socrates": 0.9963822847902685}, "confidence": 0.9834310974715067, "choice": "socrates"}}, "usage": {"input_tokens": 335, "output_tokens": 1}, "latency_ms": 185.7086039999558, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 335, "compiled_prompt_sha256": "93158928479c6a94dccd8832caae1cd3689d9b803d2128164772c4d869e1d61c", "compiled_ids_sha256": "c7fed4f81f87faf699d6216673c4908d2799b50406ff5284a7234adb84129bd2", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [13.375, 13.3125, 13.125, 13.3125, 13.75, 20.625], "runtime_probabilities": [0.0007076052133925259, 0.0006647336413152516, 0.000551083532627672, 0.0006647336413152516, 0.0010295595275238156, 0.9963823556900024], "decide_seconds": 0.18174039500081562, "probabilities_f64_minus_runtime_max_abs": 7.089973397267357e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.0006653600323528608, "sagan": 0.00031429382481417084, "socrates": 0.9973213322707658, "ibn_sina": 0.0005871781676549858, "darwin": 0.0006250479058662742, "hypatia": 0.00048678779854579614}, "confidence": 0.9872916402781384, "choice": "socrates"}}, "usage": {"input_tokens": 335, "output_tokens": 1}, "latency_ms": 185.75543399856542, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 335, "compiled_prompt_sha256": "4baed2eca5b55d49908e1c2c5bcfdc217d52427193c34234525c058ceca51e6b", "compiled_ids_sha256": "5536998ca5c3a74592ec468ce00ea29562c01a1872ec9228f1a5e095d2782f86", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [13.1875, 12.4375, 20.5, 13.0625, 13.125, 12.875], "runtime_probabilities": [0.0006653600139543414, 0.00031429383670911193, 0.9973213076591492, 0.000587178161367774, 0.0006250478327274323, 0.00048678781604394317], "decide_seconds": 0.18161631600014516, "probabilities_f64_minus_runtime_max_abs": 2.4611616633585243e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.998875685174797, "ibn_sina": 0.00020323960650690562, "hypatia": 0.00031478366044037433, "einstein": 0.0001793583232248562, "sagan": 0.00029571188257792924, "darwin": 0.00013122135245296823}, "confidence": 0.9941326707849321, "choice": "socrates"}}, "usage": {"input_tokens": 335, "output_tokens": 1}, "latency_ms": 185.5517059993872, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 335, "compiled_prompt_sha256": "53075d3ffd6a2d6a6792062761c6c2a82284c8b143a93e98e928e589e597923a", "compiled_ids_sha256": "0959a9a7eba61048e8fe4d8c7762a8dbe3dfa2b38be6b720445cb7768b90724d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.875, 12.375, 12.8125, 12.25, 12.75, 11.9375], "runtime_probabilities": [0.9988756775856018, 0.0002032396150752902, 0.0003147836832795292, 0.00017935832147486508, 0.00029571191407740116, 0.00013122135715093464], "decide_seconds": 0.18170592499882332, "probabilities_f64_minus_runtime_max_abs": 7.58919516119505e-09}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:32Z"} {"id": "c-socrates-01-starter-03", "uid": "c106", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "prompt_sha256_by_order": ["4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269"], "request_sha256": "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "compiled_prompt_sha256_by_order": ["e89581a53f340a2aeeb9384cd75a6f0aefe0ea16e3efc03c033b3b5f3f2847e8", "437efd1d26cd0c26910a20a82f906c6b9705a9e3fd974ff00769dbe6bdaf0da4", "6851eeebfdc5de3a4d20c0887d25357faaeefc62bba8c472feccf9a3baefe27f", "73cee36163244eb9d3e50cb1af31d65e1916fd5439c18ec841c3a5440bc7c96b", "305844e65da72bfb66ab061ddd0170a9fa893faa80b4c85fabf7f0079e632b50", "1421a4039fffbea0b767ede39fb42de4af16934e20c68dd7afb0522c34e5817c"], "compiled_tokens_by_order": [355, 355, 355, 355, 355, 355], "choice": "socrates", "probs": {"darwin": 0.003802266314818141, "einstein": 0.0007487106296233175, "hypatia": 0.0031521894900607235, "ibn_sina": 0.0011596257080700328, "sagan": 0.0007487106296233175, "socrates": 0.9903884972278045}, "confidence": 0.9623144862229441, "seconds": 0.18748539366667197, "seconds_all_orders": 1.1249123620000319, "seconds_by_order": [0.18763420899995253, 0.1873595610013581, 0.18710326399741462, 0.18750401000215788, 0.18769480899936752, 0.18761650899978122], "argmax_stable": true, "spread": {"darwin": 0.002892659254394727, "einstein": 0.0010279825970373776, "hypatia": 0.004717980044594105, "ibn_sina": 0.0016630873348902372, "sagan": 0.0005666046545222639, "socrates": 0.009902219559370873}, "spread_2dp": {"darwin": 0.0, "einstein": 0.0, "hypatia": 0.01, "ibn_sina": 0.0, "sagan": 0.0, "socrates": 0.010000000000000009}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.003802266314818141, "einstein": 0.0007487106296233175, "hypatia": 0.0031521894900607235, "ibn_sina": 0.0011596257080700328, "sagan": 0.0007487106296233175, "socrates": 0.9903884972278045}, {"ibn_sina": 0.0014021684455481369, "hypatia": 0.0010258483491788351, "sagan": 0.0004017279856555375, "socrates": 0.9927907611609842, "darwin": 0.0035805625609802035, "einstein": 0.0007989314976529848}, {"hypatia": 0.005516705430063569, "darwin": 0.0037915724963306307, "sagan": 0.0007013703882212724, "ibn_sina": 0.0011563642776996185, "einstein": 0.0012309433661017964, "socrates": 0.9876030440415832}, {"ibn_sina": 0.001907904753981485, "einstein": 0.0009012303915964279, "darwin": 0.0024497981967315993, "sagan": 0.0009012303915964279, "hypatia": 0.005520705457986847, "socrates": 0.9883191308081073}, {"einstein": 0.0009055241746824082, "sagan": 0.0004018239167287085, "socrates": 0.9930278357152925, "ibn_sina": 0.0011627160557174023, "darwin": 0.002461469909269564, "hypatia": 0.0020406302283093366}, {"socrates": 0.997505263600954, "ibn_sina": 0.00024481741909124796, "hypatia": 0.0008027254133927416, "einstein": 0.0002029607690644188, "sagan": 0.00033462573707416387, "darwin": 0.0009096070604234138}], "prompt_tokens": 355, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.003802266314818141, "einstein": 0.0007487106296233175, "hypatia": 0.0031521894900607235, "ibn_sina": 0.0011596257080700328, "sagan": 0.0007487106296233175, "socrates": 0.9903884972278045}, "confidence": 0.9623144862229441, "choice": "socrates"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.60591899990686, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "e89581a53f340a2aeeb9384cd75a6f0aefe0ea16e3efc03c033b3b5f3f2847e8", "compiled_ids_sha256": "67d6cefb26c2e1a4ee433d8b03c2659df0da773a4de7d34f23c8e3cc19bf6703", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.1875, 13.5625, 15.0, 14.0, 13.5625, 20.75], "runtime_probabilities": [0.0038022662047296762, 0.0007487105904147029, 0.0031521895434707403, 0.001159625593572855, 0.0007487105904147029, 0.9903884530067444], "decide_seconds": 0.1833876920027251, "probabilities_f64_minus_runtime_max_abs": 4.422106014700944e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0014021684455481369, "hypatia": 0.0010258483491788351, "sagan": 0.0004017279856555375, "socrates": 0.9927907611609842, "darwin": 0.0035805625609802035, "einstein": 0.0007989314976529848}, "confidence": 0.9707207686545941, "choice": "socrates"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.33416100076283, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "437efd1d26cd0c26910a20a82f906c6b9705a9e3fd974ff00769dbe6bdaf0da4", "compiled_ids_sha256": "9c355caacf02925357ec3acdff90bede1d61a96996c01772c656b4544891bf14", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.4375, 14.125, 13.1875, 21.0, 15.375, 13.875], "runtime_probabilities": [0.0014021684182807803, 0.0010258483234792948, 0.0004017279716208577, 0.992790699005127, 0.003580562537536025, 0.0007989314617589116], "decide_seconds": 0.1831940840020252, "probabilities_f64_minus_runtime_max_abs": 6.215585723889916e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.005516705430063569, "darwin": 0.0037915724963306307, "sagan": 0.0007013703882212724, "ibn_sina": 0.0011563642776996185, "einstein": 0.0012309433661017964, "socrates": 0.9876030440415832}, "confidence": 0.9535064275846679, "choice": "socrates"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.07843399897683, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "6851eeebfdc5de3a4d20c0887d25357faaeefc62bba8c472feccf9a3baefe27f", "compiled_ids_sha256": "2dbadb8722bbc6830cf4d7deec36aef1e8c4e463d32a1d87397cba3151d349b6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.4375, 15.0625, 13.375, 13.875, 13.9375, 20.625], "runtime_probabilities": [0.00551670603454113, 0.003791572991758585, 0.0007013704162091017, 0.0011563643347471952, 0.0012309434823691845, 0.9876031279563904], "decide_seconds": 0.18292425599793205, "probabilities_f64_minus_runtime_max_abs": 8.39148072140361e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.001907904753981485, "einstein": 0.0009012303915964279, "darwin": 0.0024497981967315993, "sagan": 0.0009012303915964279, "hypatia": 0.005520705457986847, "socrates": 0.9883191308081073}, "confidence": 0.9555583733168691, "choice": "socrates"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.478760999511, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "73cee36163244eb9d3e50cb1af31d65e1916fd5439c18ec841c3a5440bc7c96b", "compiled_ids_sha256": "430b0f5076d8b2a4d0f731595846a007bd4023e82a914e182af6e9f23bb6f11d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [14.5, 13.75, 14.75, 13.75, 15.5625, 20.75], "runtime_probabilities": [0.0019079050980508327, 0.0009012305526994169, 0.0024497986305505037, 0.0009012305526994169, 0.005520706530660391, 0.9883192777633667], "decide_seconds": 0.183240113001375, "probabilities_f64_minus_runtime_max_abs": 1.469552594057788e-07}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.0009055241746824082, "sagan": 0.0004018239167287085, "socrates": 0.9930278357152925, "ibn_sina": 0.0011627160557174023, "darwin": 0.002461469909269564, "hypatia": 0.0020406302283093366}, "confidence": 0.9711356042924317, "choice": "socrates"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.66898899775697, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "305844e65da72bfb66ab061ddd0170a9fa893faa80b4c85fabf7f0079e632b50", "compiled_ids_sha256": "d32001222704ea3f186946ce10cb8bea410d7009ed7f76e50a1bac7f479b0b17", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [13.875, 13.0625, 20.875, 14.125, 14.875, 14.6875], "runtime_probabilities": [0.000905524124391377, 0.0004018238978460431, 0.9930277466773987, 0.0011627159547060728, 0.0024614697322249413, 0.0020406299736350775], "decide_seconds": 0.18335145200035186, "probabilities_f64_minus_runtime_max_abs": 8.903789383563065e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.997505263600954, "ibn_sina": 0.00024481741909124796, "hypatia": 0.0008027254133927416, "einstein": 0.0002029607690644188, "sagan": 0.00033462573707416387, "darwin": 0.0009096070604234138}, "confidence": 0.9882675442639713, "choice": "socrates"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 187.5912600007723, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 355, "compiled_prompt_sha256": "1421a4039fffbea0b767ede39fb42de4af16934e20c68dd7afb0522c34e5817c", "compiled_ids_sha256": "6c2ed8eba018f460b59ea195da78e10ed5ae741cc48d9199f842e9d0bfe26c6f", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.875, 12.5625, 13.75, 12.375, 12.875, 13.875], "runtime_probabilities": [0.997505247592926, 0.00024481743457727134, 0.0008027253788895905, 0.00020296077127568424, 0.0003346257144585252, 0.0009096070425584912], "decide_seconds": 0.18345673099975102, "probabilities_f64_minus_runtime_max_abs": 1.600802801426937e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:33Z"} {"id": "c-socrates-01-starter-03", "uid": "c107", "task": "kperm", "label": "socrates", "mode": "apus-six-kev-orders", "n_options": 6, "n_perm": 6, "prompt_sha256": "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "prompt_sha256_by_order": ["0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296"], "request_sha256": "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "compiled_prompt_sha256_by_order": ["8501d29a2806beb81cce7eb55e95f965f1f81dc43c69104f947a1d2360bdb71d", "77c3f008f54260781cc6a5a98c14bdcded59801edeea2bf7da2cfd556c6b7e46", "f7a4bd321ebb69865fae5b8e39d55a92b75a7c098cc7a6436e526e0b8b977f30", "1ba1a9e9577f8ba56c3d1708f86214fc84474ba1da0d1f66f9939b6f8302490d", "f3ac6ac3c0fe3d8ed813c3e686bf2b9748592aaf359591a7c06bbb432feef8bd", "4b1c426f4500d6c29cc62fe4b80b73d9f8664fa34b037279b6feb205538d4de8"], "compiled_tokens_by_order": [328, 328, 328, 328, 328, 328], "choice": "socrates", "probs": {"darwin": 0.026380209677729822, "einstein": 0.018130835283761554, "hypatia": 0.033872859723757025, "ibn_sina": 0.018130835283761554, "sagan": 0.029892693787898064, "socrates": 0.8735925662430919}, "confidence": 0.6768759779057689, "seconds": 0.1847325485014153, "seconds_all_orders": 1.1083952910084918, "seconds_by_order": [0.18495142000028864, 0.1843730950022291, 0.18447047399968142, 0.18510177900316194, 0.1849054000012984, 0.18459312300183228], "argmax_stable": true, "spread": {"darwin": 0.02763478705193289, "einstein": 0.035952331837798016, "hypatia": 0.02973038980187124, "ibn_sina": 0.018612812118384434, "sagan": 0.03036729011367184, "socrates": 0.12520733802029527}, "spread_2dp": {"darwin": 0.03, "einstein": 0.04, "hypatia": 0.03, "ibn_sina": 0.01, "sagan": 0.03, "socrates": 0.13}, "choices_by_order": ["socrates", "socrates", "socrates", "socrates", "socrates", "socrates"], "orders": [["darwin", "einstein", "hypatia", "ibn_sina", "sagan", "socrates"], ["ibn_sina", "hypatia", "sagan", "socrates", "darwin", "einstein"], ["hypatia", "darwin", "sagan", "ibn_sina", "einstein", "socrates"], ["ibn_sina", "einstein", "darwin", "sagan", "hypatia", "socrates"], ["einstein", "sagan", "socrates", "ibn_sina", "darwin", "hypatia"], ["socrates", "ibn_sina", "hypatia", "einstein", "sagan", "darwin"]], "probabilities_by_order": [{"darwin": 0.026380209677729822, "einstein": 0.018130835283761554, "hypatia": 0.033872859723757025, "ibn_sina": 0.018130835283761554, "sagan": 0.029892693787898064, "socrates": 0.8735925662430919}, {"ibn_sina": 0.0137255902190133, "hypatia": 0.006483509736091101, "sagan": 0.027296595560683964, "socrates": 0.9039390989256837, "darwin": 0.027296595560683964, "einstein": 0.021258609997844104}, {"hypatia": 0.0206511810233711, "darwin": 0.01608315595235075, "sagan": 0.03858148546311409, "ibn_sina": 0.01252557444995029, "einstein": 0.034048041418310776, "socrates": 0.878110561692903}, {"ibn_sina": 0.02488942489565486, "einstein": 0.04649956744412011, "darwin": 0.03621389953796234, "sagan": 0.031958654172761475, "hypatia": 0.03621389953796234, "socrates": 0.824224554411539}, {"einstein": 0.010547235606322088, "sagan": 0.008214195349442246, "socrates": 0.9494318924318342, "ibn_sina": 0.013542918594311504, "darwin": 0.010547235606322088, "hypatia": 0.007716522411767749}, {"socrates": 0.9315319307383011, "ibn_sina": 0.006276612777270427, "hypatia": 0.01706160245672806, "einstein": 0.011726256447502909, "sagan": 0.024824485094168142, "darwin": 0.008579112486029454}], "prompt_tokens": 328, "eval_tokens": 0, "apus_responses": [{"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.026380209677729822, "einstein": 0.018130835283761554, "hypatia": 0.033872859723757025, "ibn_sina": 0.018130835283761554, "sagan": 0.029892693787898064, "socrates": 0.8735925662430919}, "confidence": 0.6768759779057689, "choice": "socrates"}}, "usage": {"input_tokens": 328, "output_tokens": 1}, "latency_ms": 184.92547000278137, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 328, "compiled_prompt_sha256": "8501d29a2806beb81cce7eb55e95f965f1f81dc43c69104f947a1d2360bdb71d", "compiled_ids_sha256": "da1348b4fbd3b223210453e72685938201ef08da043c35159863b5490ea36eae", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.625, 16.25, 16.875, 16.25, 16.75, 20.125], "runtime_probabilities": [0.026380209252238274, 0.0181308351457119, 0.03387285768985748, 0.0181308351457119, 0.02989269234240055, 0.8735925555229187], "decide_seconds": 0.18091822199858143, "probabilities_f64_minus_runtime_max_abs": 1.0720173193590199e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.0137255902190133, "hypatia": 0.006483509736091101, "sagan": 0.027296595560683964, "socrates": 0.9039390989256837, "darwin": 0.027296595560683964, "einstein": 0.021258609997844104}, "confidence": 0.7425561343432665, "choice": "socrates"}}, "usage": {"input_tokens": 328, "output_tokens": 1}, "latency_ms": 184.35040499753086, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 328, "compiled_prompt_sha256": "77c3f008f54260781cc6a5a98c14bdcded59801edeea2bf7da2cfd556c6b7e46", "compiled_ids_sha256": "b0649dc602a40b6a3aef2c126bbaff7cfd5ed4d4f14f93fc364d3a925a16cffb", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [15.9375, 15.1875, 16.625, 20.125, 16.625, 16.375], "runtime_probabilities": [0.013725590892136097, 0.006483509670943022, 0.027296599000692368, 0.9039391875267029, 0.027296599000692368, 0.021258613094687462], "decide_seconds": 0.18047016499986057, "probabilities_f64_minus_runtime_max_abs": 8.860101918806151e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"hypatia": 0.0206511810233711, "darwin": 0.01608315595235075, "sagan": 0.03858148546311409, "ibn_sina": 0.01252557444995029, "einstein": 0.034048041418310776, "socrates": 0.878110561692903}, "confidence": 0.6895708943700112, "choice": "socrates"}}, "usage": {"input_tokens": 328, "output_tokens": 1}, "latency_ms": 184.4480539984943, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 328, "compiled_prompt_sha256": "f7a4bd321ebb69865fae5b8e39d55a92b75a7c098cc7a6436e526e0b8b977f30", "compiled_ids_sha256": "cbca5d517839ffd7217054ee63cbe830dcbc5074db77f7af2b4710b3ac8580a6", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 16.25, 17.125, 16.0, 17.0, 20.25], "runtime_probabilities": [0.020651182159781456, 0.016083156690001488, 0.0385814867913723, 0.012525574304163456, 0.03404804319143295, 0.8781105875968933], "decide_seconds": 0.18052032500054338, "probabilities_f64_minus_runtime_max_abs": 2.5903990286124667e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"ibn_sina": 0.02488942489565486, "einstein": 0.04649956744412011, "darwin": 0.03621389953796234, "sagan": 0.031958654172761475, "hypatia": 0.03621389953796234, "socrates": 0.824224554411539}, "confidence": 0.5845904412939255, "choice": "socrates"}}, "usage": {"input_tokens": 328, "output_tokens": 1}, "latency_ms": 185.07899900214397, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 328, "compiled_prompt_sha256": "1ba1a9e9577f8ba56c3d1708f86214fc84474ba1da0d1f66f9939b6f8302490d", "compiled_ids_sha256": "09b6aaff162d3e86d4434e1e1271911d9cda6a72c822905bfb17f7460e720e88", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.5, 17.125, 16.875, 16.75, 16.875, 20.0], "runtime_probabilities": [0.024889424443244934, 0.04649956896901131, 0.03621390089392662, 0.03195865452289581, 0.03621390089392662, 0.8242245316505432], "decide_seconds": 0.18119250899690087, "probabilities_f64_minus_runtime_max_abs": 2.276099575482249e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"einstein": 0.010547235606322088, "sagan": 0.008214195349442246, "socrates": 0.9494318924318342, "ibn_sina": 0.013542918594311504, "darwin": 0.010547235606322088, "hypatia": 0.007716522411767749}, "confidence": 0.843434789909503, "choice": "socrates"}}, "usage": {"input_tokens": 328, "output_tokens": 1}, "latency_ms": 184.88233999960357, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 328, "compiled_prompt_sha256": "f3ac6ac3c0fe3d8ed813c3e686bf2b9748592aaf359591a7c06bbb432feef8bd", "compiled_ids_sha256": "ee35be70936ae84929a015fe734388ecdf428920a8ec620d93275f85402d2351", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [16.125, 15.875, 20.625, 16.375, 16.125, 15.8125], "runtime_probabilities": [0.010547236539423466, 0.008214195258915424, 0.9494319558143616, 0.013542918488383293, 0.010547236539423466, 0.0077165234833955765], "decide_seconds": 0.18082508299994515, "probabilities_f64_minus_runtime_max_abs": 6.338252733772975e-08}}}, {"model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "answers": {"q": {"type": "choice", "probabilities": {"socrates": 0.9315319307383011, "ibn_sina": 0.006276612777270427, "hypatia": 0.01706160245672806, "einstein": 0.011726256447502909, "sagan": 0.024824485094168142, "darwin": 0.008579112486029454}, "confidence": 0.8035115702462929, "choice": "socrates"}}, "usage": {"input_tokens": 328, "output_tokens": 1}, "latency_ms": 184.570002998953, "apus": {"q": {"effort": "high", "executed_layers": 32, "projection": "full_head", "calibrated": false, "primitive": "choice", "compiled_tokens": 328, "compiled_prompt_sha256": "4b1c426f4500d6c29cc62fe4b80b73d9f8664fa34b037279b6feb205538d4de8", "compiled_ids_sha256": "c50f8660ccb0895cb05200668ded6977ef98c5a3ec2f592ea50243ce1d18e38d", "candidate_token_ids": [32, 33, 34, 35, 36, 37], "logits": [20.375, 15.375, 16.375, 16.0, 16.75, 15.6875], "runtime_probabilities": [0.9315319061279297, 0.006276612635701895, 0.01706160232424736, 0.0117262564599514, 0.024824485182762146, 0.008579112589359283], "decide_seconds": 0.18069251300039468, "probabilities_f64_minus_runtime_max_abs": 2.4610371407440823e-08}}}], "orders_source": "Kev-9B's own kperm rows (its server's seeded draw; Kev-4B's are identical)", "prompt_sha256_matches_kev9b": true, "card": "RTX 3090 (fp 33a0b4acb8e8)", "box": "benchbox", "os": {"pretty_name": "Ubuntu 26.04 LTS", "id": "ubuntu", "version_id": "26.04", "kernel": "7.0.0-31-generic", "arch": "x86_64", "source": "/etc/os-release"}, "dry": false, "preflight": false, "effort": "high", "runtime": "openjet_runtime (APUS-OpenJev-v1-9B@82c9c56c) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "82c9c56cfa9de8d36704ed91948d4726ef111635", "bridge_revision": "af276da53a2e7a6d0fbdbafc9280bf9982b2d3a6", "torch": "2.14.0+cu130", "torch_cuda": "13.0", "transformers": "5.16.1", "tokenizers": "0.23.2", "dtype": "bfloat16", "kernels": "reference (no flash-linear-attention), sdpa", "calibrated": false, "arm": "apus-9b-high", "model": "apus-ailab/APUS-OpenJev-v1-9B@82c9c56c", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T15:38:34Z"}