{"mode": "kev-choice", "prompt_sha256": "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "n_options": 6, "kev_response": {"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": 206.62269199965522, "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": 0.19995597299930523, "probabilities_f64_minus_runtime_max_abs": 2.2846654623709384e-08}}}, "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": 0.2066795809987525, "prompt_tokens": 279, "eval_tokens": 1, "server_latency_ms": 206.62269199965522, "prompt_eval_ms": null, "total_ms": 206.6795809987525, "done_reason": null, "id": "c-darwin-01-dessert-01", "uid": "c000", "task": "kc", "label": "darwin", "text_sha256": "fcf5cde7b44af0c94fc1aa5d71cb854130f81c4f17c238bcf352c844972bb393", "request_sha256": "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "compiled_tokens": 279, "compiled_prompt_sha256": "2ca65940a5765f77af59cf411c0b326a074bbc7804d92041b9a96860a2ba21f0", "compiled_ids_sha256": "39875e59687c95458e3c8bef14fa71362355147df3bbbb6895cd7ad2bb135fb8", "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": "2026-09-28T15:32:39Z"} {"mode": "kev-choice", "prompt_sha256": "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "n_options": 6, "kev_response": {"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": 190.90081399917835, "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": 0.18640335899908678, "probabilities_f64_minus_runtime_max_abs": 6.12855621806574e-08}}}, "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": 0.19092864299818757, "prompt_tokens": 379, "eval_tokens": 1, "server_latency_ms": 190.90081399917835, "prompt_eval_ms": null, "total_ms": 190.92864299818757, "done_reason": null, "id": "c-darwin-01-dessert-01", "uid": "c001", "task": "kc", "label": "darwin", "text_sha256": "cca09c399a86e9ba18f164d04a4ceff22429e46d782c68ba7651b4f8f03254c9", "request_sha256": "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "compiled_tokens": 379, "compiled_prompt_sha256": "dfdf84665a6deae6826cfe5cb74cbdf7164c3c10210051c62958077957da1655", "compiled_ids_sha256": "935f8d246f67d04de94e49c25e2c59f166760a7b2d6dddf4f9fb2ac6ee141755", "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": "2026-09-28T15:32:39Z"} {"mode": "kev-choice", "prompt_sha256": "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "n_options": 6, "kev_response": {"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": 170.98331899978803, "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": 0.16742499599786242, "probabilities_f64_minus_runtime_max_abs": 9.86362322374923e-09}}}, "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": 0.171007708999241, "prompt_tokens": 296, "eval_tokens": 1, "server_latency_ms": 170.98331899978803, "prompt_eval_ms": null, "total_ms": 171.007708999241, "done_reason": null, "id": "c-darwin-01-dessert-01", "uid": "c002", "task": "kc", "label": "darwin", "text_sha256": "a062080c339f6ce54cbabf2848b449ef14a2780b84a5f4583d04f0156553982d", "request_sha256": "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "compiled_tokens": 296, "compiled_prompt_sha256": "36256d9ffe4d50c3d22a8cefe2d22c38863f669314f6c78f6e3165edd736e631", "compiled_ids_sha256": "bedc1945b96a6e27ccb028c42f119ba2eb225cdab520d6ff987d956978f024cd", "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": "2026-09-28T15:32:39Z"} {"mode": "kev-choice", "prompt_sha256": "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "n_options": 6, "kev_response": {"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": 205.8568379979988, "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": 0.20130081299794256, "probabilities_f64_minus_runtime_max_abs": 7.343090524525309e-08}}}, "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": 0.20588308700098423, "prompt_tokens": 399, "eval_tokens": 1, "server_latency_ms": 205.8568379979988, "prompt_eval_ms": null, "total_ms": 205.88308700098423, "done_reason": null, "id": "c-darwin-01-dessert-02", "uid": "c003", "task": "kc", "label": "darwin", "text_sha256": "485359cc7a1ff70871cdb0ae47661b859bb9bcda6d35a4782bd26dad348398ce", "request_sha256": "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "compiled_tokens": 399, "compiled_prompt_sha256": "604afd02ea457c03d0d85cec823cbc75b5a82d677e33fef5295394b9bc38e733", "compiled_ids_sha256": "77f997cd82656949ba572af316c23fd06bc92bfc51d6b9880262378e26c5a2fd", "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": "2026-09-28T15:32:39Z"} {"mode": "kev-choice", "prompt_sha256": "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "n_options": 6, "kev_response": {"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": 206.98653799991007, "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": 0.20231873499869835, "probabilities_f64_minus_runtime_max_abs": 4.19422075403908e-08}}}, "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": 0.20701312800156302, "prompt_tokens": 409, "eval_tokens": 1, "server_latency_ms": 206.98653799991007, "prompt_eval_ms": null, "total_ms": 207.01312800156302, "done_reason": null, "id": "c-darwin-01-dessert-03", "uid": "c004", "task": "kc", "label": "darwin", "text_sha256": "a059154398302a58169460b1f4228781795950865827925fc503638efc0b8cfb", "request_sha256": "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "compiled_tokens": 409, "compiled_prompt_sha256": "f23455f0797804d5728aeff3fdc9d48c4b31d4a7b4b5aa74d195e7140ad387f6", "compiled_ids_sha256": "2fc9c2e9c165a1094386aba821c96f2b186d61c9c48cddddeb2a7985bc7f809a", "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": "2026-09-28T15:32:40Z"} {"mode": "kev-choice", "prompt_sha256": "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "n_options": 6, "kev_response": {"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": 208.68582500042976, "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": 0.20382842300023185, "probabilities_f64_minus_runtime_max_abs": 1.571577046277639e-08}}}, "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": 0.2087125450016174, "prompt_tokens": 411, "eval_tokens": 1, "server_latency_ms": 208.68582500042976, "prompt_eval_ms": null, "total_ms": 208.7125450016174, "done_reason": null, "id": "c-darwin-01-dessert-03", "uid": "c005", "task": "kc", "label": "darwin", "text_sha256": "23d3cdf830a1ee0f835507f044275131684307cb945a80c51538c7b409534789", "request_sha256": "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "compiled_tokens": 411, "compiled_prompt_sha256": "24e4de0b01d855575257e46ab574c911b12ec716af49ef43360d71b0ceb39554", "compiled_ids_sha256": "7edad3a35e2849ccf2f8adce479f5cd518ba5078d299c95eded9c3ee2725b892", "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": "2026-09-28T15:32:40Z"} {"mode": "kev-choice", "prompt_sha256": "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "n_options": 6, "kev_response": {"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": 172.81879399888567, "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": 0.16893033499945886, "probabilities_f64_minus_runtime_max_abs": 2.9922080591049394e-08}}}, "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": 0.17284516400104621, "prompt_tokens": 314, "eval_tokens": 1, "server_latency_ms": 172.81879399888567, "prompt_eval_ms": null, "total_ms": 172.84516400104621, "done_reason": null, "id": "c-darwin-01-main-01", "uid": "c006", "task": "kc", "label": "darwin", "text_sha256": "5dd7557e7f11609ac3d7c3e8fc7175298f308905f62b24085bbccf753b83199d", "request_sha256": "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "compiled_tokens": 314, "compiled_prompt_sha256": "377f0fd9228b401ffd1752b4418d7a4f949f5c9987b66900bf4451b634fa3b16", "compiled_ids_sha256": "01e1272f80d0db6c2b279d6449fa670109af77f7a80b594d0041b4894a23f3be", "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": "2026-09-28T15:32:40Z"} {"mode": "kev-choice", "prompt_sha256": "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "n_options": 6, "kev_response": {"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": 170.20299500291003, "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": 0.1667604520007444, "probabilities_f64_minus_runtime_max_abs": 1.0146878309802432e-08}}}, "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": 0.17022677500062855, "prompt_tokens": 295, "eval_tokens": 1, "server_latency_ms": 170.20299500291003, "prompt_eval_ms": null, "total_ms": 170.22677500062855, "done_reason": null, "id": "c-darwin-01-main-02", "uid": "c007", "task": "kc", "label": "darwin", "text_sha256": "e5582302f777af6b5f897c4ab2cf415650f9c356ad9b79067cf05fd8ebb156d8", "request_sha256": "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "compiled_tokens": 295, "compiled_prompt_sha256": "98b6551eb87406f35159db42e99ab68b6616d9349e25272937a35b9bb812bc44", "compiled_ids_sha256": "1e5073f8ebc6dc8593ad0d9a35127317ad2b076509d4a0a08bc1f851b7949344", "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": "2026-09-28T15:32:40Z"} {"mode": "kev-choice", "prompt_sha256": "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "n_options": 6, "kev_response": {"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": 189.62758400084567, "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": 0.18541721599831362, "probabilities_f64_minus_runtime_max_abs": 2.2105606012345902e-08}}}, "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": 0.18965988300260506, "prompt_tokens": 372, "eval_tokens": 1, "server_latency_ms": 189.62758400084567, "prompt_eval_ms": null, "total_ms": 189.65988300260506, "done_reason": null, "id": "c-darwin-01-main-02", "uid": "c008", "task": "kc", "label": "darwin", "text_sha256": "649c28612e897217908c8e2d4d59f51856253ea1c610f2c4ce6cd466b628185e", "request_sha256": "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "compiled_tokens": 372, "compiled_prompt_sha256": "a94c81da04c23301082578b21ecddd1e9b9cf59dce663b78a8e4f45435535a31", "compiled_ids_sha256": "6aa68021cc66abd7749c997fa9d95159d904c84b3398aad7280decc31822aa57", "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": "2026-09-28T15:32:40Z"} {"mode": "kev-choice", "prompt_sha256": "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "n_options": 6, "kev_response": {"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": 207.78258299833396, "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": 0.20319554800153128, "probabilities_f64_minus_runtime_max_abs": 7.80366725683379e-08}}}, "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": 0.2078083129999868, "prompt_tokens": 407, "eval_tokens": 1, "server_latency_ms": 207.78258299833396, "prompt_eval_ms": null, "total_ms": 207.8083129999868, "done_reason": null, "id": "c-darwin-01-main-02", "uid": "c009", "task": "kc", "label": "darwin", "text_sha256": "5d08013bae51d2d1f6280a290baf0b7fbc1b4a3cd4f743a5cb23ab10f96bee6a", "request_sha256": "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "compiled_tokens": 407, "compiled_prompt_sha256": "67d1a2d7e7aebb688c009d8c350b3caab3a77e09611db29376cf5a84ebb7f2f5", "compiled_ids_sha256": "b6064067571c6aaaef5c497a87a07eae439d78eab1c8465b4f06fd5fd0319cf5", "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": "2026-09-28T15:32:40Z"} {"mode": "kev-choice", "prompt_sha256": "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "n_options": 6, "kev_response": {"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": 206.2083739983791, "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": 0.20164416099942173, "probabilities_f64_minus_runtime_max_abs": 6.979805677875817e-09}}}, "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": 0.2062362039978325, "prompt_tokens": 401, "eval_tokens": 1, "server_latency_ms": 206.2083739983791, "prompt_eval_ms": null, "total_ms": 206.2362039978325, "done_reason": null, "id": "c-darwin-01-main-02", "uid": "c010", "task": "kc", "label": "darwin", "text_sha256": "51ea227691d5922785fa53ce48e8f0d2c1e132ed64a61475fabf1b25d10857d8", "request_sha256": "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "compiled_tokens": 401, "compiled_prompt_sha256": "845aaf3e77ec7fd9ccefeb0369bad6914f74a0a4e47c5038bcac81310111f0a8", "compiled_ids_sha256": "2c82207f9a93c3dc26bdba65b250a344432061dc02ca1018c708ad8067c3121f", "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": "2026-09-28T15:32:41Z"} {"mode": "kev-choice", "prompt_sha256": "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "n_options": 6, "kev_response": {"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": 207.53345399862155, "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": 0.20288014100151486, "probabilities_f64_minus_runtime_max_abs": 8.143881458089197e-09}}}, "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": 0.20755985399955534, "prompt_tokens": 402, "eval_tokens": 1, "server_latency_ms": 207.53345399862155, "prompt_eval_ms": null, "total_ms": 207.55985399955534, "done_reason": null, "id": "c-darwin-01-main-03", "uid": "c011", "task": "kc", "label": "darwin", "text_sha256": "460ae9db9052213652d36735be593ecd3ddee733eb946607b120386c23188fc5", "request_sha256": "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "compiled_tokens": 402, "compiled_prompt_sha256": "67ed85f77269ae70491a1864dafa1c40fd5c3c6a2590a97dc9bb1b6af2efd7d9", "compiled_ids_sha256": "f12e7fc6a162cdd7569eba4da42ac19c958d6ecfbd2351179f1c50f8bd12014a", "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": "2026-09-28T15:32:41Z"} {"mode": "kev-choice", "prompt_sha256": "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "n_options": 6, "kev_response": {"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": 160.62135899846908, "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": 0.1576957229990512, "probabilities_f64_minus_runtime_max_abs": 3.736206444227008e-08}}}, "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": 0.16064290899885236, "prompt_tokens": 240, "eval_tokens": 1, "server_latency_ms": 160.62135899846908, "prompt_eval_ms": null, "total_ms": 160.64290899885236, "done_reason": null, "id": "c-darwin-01-starter-01", "uid": "c012", "task": "kc", "label": "darwin", "text_sha256": "6847f4760fc0710fb9f00b2fdee4feda7fb17ee4a2e7a43cc1c094926b2995d6", "request_sha256": "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "compiled_tokens": 240, "compiled_prompt_sha256": "9e3f59abe3f3521d1414d75abb72215d06b06fabcaa7452231957e31bc3997a0", "compiled_ids_sha256": "79ebf5af4c3e548555cf057c1040dde2bbc5c096909a258efb54835918c41f5f", "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": "2026-09-28T15:32:41Z"} {"mode": "kev-choice", "prompt_sha256": "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "n_options": 6, "kev_response": {"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": 169.03965399978915, "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": 0.16580147899730946, "probabilities_f64_minus_runtime_max_abs": 6.049689449394435e-08}}}, "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": 0.16906398399805767, "prompt_tokens": 269, "eval_tokens": 1, "server_latency_ms": 169.03965399978915, "prompt_eval_ms": null, "total_ms": 169.06398399805767, "done_reason": null, "id": "c-darwin-01-starter-01", "uid": "c013", "task": "kc", "label": "darwin", "text_sha256": "5abf6677c76d7e1ad5b7e964a4153f41b1c6ea65102069c921a1fbb1ee1615de", "request_sha256": "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "compiled_tokens": 269, "compiled_prompt_sha256": "4503591e6f0e5386c6e8577056733fa060b0d80509ece075b927a593b418dd73", "compiled_ids_sha256": "fc3048d46a00e63f863e4d4d211b3e8a6c309b8b184a6adb0f036ee0575056f9", "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": "2026-09-28T15:32:41Z"} {"mode": "kev-choice", "prompt_sha256": "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "n_options": 6, "kev_response": {"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": 170.85735000000568, "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": 0.16738439699838636, "probabilities_f64_minus_runtime_max_abs": 2.2670053556606007e-08}}}, "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": 0.17088101000263123, "prompt_tokens": 294, "eval_tokens": 1, "server_latency_ms": 170.85735000000568, "prompt_eval_ms": null, "total_ms": 170.88101000263123, "done_reason": null, "id": "c-darwin-01-starter-01", "uid": "c014", "task": "kc", "label": "darwin", "text_sha256": "9dba6878c1f2992e4e0050abb3696a985e529b9b3ccf8a4dbb61b3fdfd630887", "request_sha256": "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "compiled_tokens": 294, "compiled_prompt_sha256": "5d90f70b3e510382c9f8fe500bb3610f93699a121cbbf98eabe2725ae32a0559", "compiled_ids_sha256": "1360cb8f139ab7df74cce524630cc666c0ab3b9f2b63c789ee0517529cc2e1d3", "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": "2026-09-28T15:32:41Z"} {"mode": "kev-choice", "prompt_sha256": "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "n_options": 6, "kev_response": {"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": 168.83964499720605, "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": 0.16568283000015072, "probabilities_f64_minus_runtime_max_abs": 1.2762265827470642e-08}}}, "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": 0.16886170499856235, "prompt_tokens": 273, "eval_tokens": 1, "server_latency_ms": 168.83964499720605, "prompt_eval_ms": null, "total_ms": 168.86170499856235, "done_reason": null, "id": "c-darwin-01-starter-02", "uid": "c015", "task": "kc", "label": "darwin", "text_sha256": "e23a934fa1ab74ba1d2ea29881d37ac5294c3113e530747a58b65988ccce7b3f", "request_sha256": "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "compiled_tokens": 273, "compiled_prompt_sha256": "cc1df5806cd5c60ea952970a2eda3feb0945e82bb626caee1c5af7534e8df1ee", "compiled_ids_sha256": "f791a198d080ca74c8294b34a536763716401dc9291fa42c4a1acd27960ce18f", "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": "2026-09-28T15:32:42Z"} {"mode": "kev-choice", "prompt_sha256": "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "n_options": 6, "kev_response": {"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": 207.28157600024133, "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": 0.20279324100192753, "probabilities_f64_minus_runtime_max_abs": 1.2332606824649872e-08}}}, "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": 0.20730761600134429, "prompt_tokens": 396, "eval_tokens": 1, "server_latency_ms": 207.28157600024133, "prompt_eval_ms": null, "total_ms": 207.3076160013443, "done_reason": null, "id": "c-darwin-01-starter-03", "uid": "c016", "task": "kc", "label": "darwin", "text_sha256": "ed21081e080015da7bfcd7eab39062843beed6a7101b806f582d6b920b521b74", "request_sha256": "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "compiled_tokens": 396, "compiled_prompt_sha256": "48c20dba52dff386d2809ed467785e33fb1eac6e82e19292cd9497781a43ff1b", "compiled_ids_sha256": "ec6e841b4247aaa12db66e4d4109704bc2a69c4a9bbb1d5a4325d9879d6dca16", "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": "2026-09-28T15:32:42Z"} {"mode": "kev-choice", "prompt_sha256": "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "n_options": 6, "kev_response": {"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": 192.5299009999435, "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": 0.1881938349979464, "probabilities_f64_minus_runtime_max_abs": 6.182486633754536e-09}}}, "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": 0.19255815100041218, "prompt_tokens": 383, "eval_tokens": 1, "server_latency_ms": 192.5299009999435, "prompt_eval_ms": null, "total_ms": 192.55815100041218, "done_reason": null, "id": "c-darwin-01-starter-03", "uid": "c017", "task": "kc", "label": "darwin", "text_sha256": "62da45e279ad99e725097784fd077fdc2c79b7c185dfcddc965fc9cf95871589", "request_sha256": "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "compiled_tokens": 383, "compiled_prompt_sha256": "292b2a1853810df91bacbfc1e9d9b69a191c104ded5e174b75e99205cda89199", "compiled_ids_sha256": "80f60e0dbaddd871d7e686f5baae1c6e910fcda7924288511f08aec576cdfe5e", "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": "2026-09-28T15:32:42Z"} {"mode": "kev-choice", "prompt_sha256": "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "n_options": 6, "kev_response": {"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": 191.21670100139454, "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": 0.18679734599936637, "probabilities_f64_minus_runtime_max_abs": 1.1631684015966925e-07}}}, "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": 0.19124487099907128, "prompt_tokens": 373, "eval_tokens": 1, "server_latency_ms": 191.21670100139454, "prompt_eval_ms": null, "total_ms": 191.24487099907128, "done_reason": null, "id": "c-einstein-01-dessert-01", "uid": "c018", "task": "kc", "label": "einstein", "text_sha256": "b781300618d56e5dcd736c459a144267a9267ffc8d98516bbbb276ca5d4718af", "request_sha256": "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "compiled_tokens": 373, "compiled_prompt_sha256": "511f397ccbe1d23a7056a582c58d20ec2dc37e7b408ccc37042cd3604b440fdb", "compiled_ids_sha256": "95907c564f0990963bab66233f57176e924cae32017e3e77c3639ed1e429d350", "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": "2026-09-28T15:32:42Z"} {"mode": "kev-choice", "prompt_sha256": "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "n_options": 6, "kev_response": {"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": 191.784197002562, "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": 0.18747453000105452, "probabilities_f64_minus_runtime_max_abs": 5.1794337041499716e-08}}}, "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": 0.19181536700125434, "prompt_tokens": 380, "eval_tokens": 1, "server_latency_ms": 191.784197002562, "prompt_eval_ms": null, "total_ms": 191.81536700125434, "done_reason": null, "id": "c-einstein-01-dessert-01", "uid": "c019", "task": "kc", "label": "einstein", "text_sha256": "eb84fb712ad2783d41fceb39dc694dbb9cf5eb13f70b5718420c281b9e7774de", "request_sha256": "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "compiled_tokens": 380, "compiled_prompt_sha256": "482a99aea3518461af305b2fde5179c25aa0874e22f2f00e33c301df075be35a", "compiled_ids_sha256": "0d8c04b699286490547a95aa4750df3f99b6a6ff4c2170fbcda9228d7de90d1c", "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": "2026-09-28T15:32:42Z"} {"mode": "kev-choice", "prompt_sha256": "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "n_options": 6, "kev_response": {"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": 183.06209400179796, "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": 0.17913328600116074, "probabilities_f64_minus_runtime_max_abs": 1.9536899986682954e-08}}}, "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": 0.18308538400015095, "prompt_tokens": 349, "eval_tokens": 1, "server_latency_ms": 183.06209400179796, "prompt_eval_ms": null, "total_ms": 183.08538400015095, "done_reason": null, "id": "c-einstein-01-dessert-02", "uid": "c020", "task": "kc", "label": "einstein", "text_sha256": "4fb0a9de68aead4be7487fe269ed3b70a9895545d5104cbd279be928e1065597", "request_sha256": "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "compiled_tokens": 349, "compiled_prompt_sha256": "bf3f0bcb078d10695d6d82ba348cba3a8f6b4084459aec08e47402eef4080ca8", "compiled_ids_sha256": "2f7bcc4bbd9219b9afcae08151a8d821af356541e49f69a31dd887caf3fb2636", "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": "2026-09-28T15:32:43Z"} {"mode": "kev-choice", "prompt_sha256": "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "n_options": 6, "kev_response": {"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": 170.55370200250763, "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": 0.1672591280002962, "probabilities_f64_minus_runtime_max_abs": 5.652166623626442e-09}}}, "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": 0.1705761720004375, "prompt_tokens": 287, "eval_tokens": 1, "server_latency_ms": 170.55370200250763, "prompt_eval_ms": null, "total_ms": 170.5761720004375, "done_reason": null, "id": "c-einstein-01-dessert-02", "uid": "c021", "task": "kc", "label": "einstein", "text_sha256": "ddfc77297962cfaed22c1394076f76a103da218260e895d72255351e20d3c1a7", "request_sha256": "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "compiled_tokens": 287, "compiled_prompt_sha256": "dc15cfd6dc8dccc74630a73d746dd43d7892002655d25aa889cf89abbcc1d5b6", "compiled_ids_sha256": "770cba00074ef91bbf9b863248bc0cf872ece1572ae250a4ff112e7036365292", "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": "2026-09-28T15:32:43Z"} {"mode": "kev-choice", "prompt_sha256": "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "n_options": 6, "kev_response": {"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": 208.0321499997808, "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": 0.20355863500299165, "probabilities_f64_minus_runtime_max_abs": 2.4952928612265168e-08}}}, "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": 0.20805832999758422, "prompt_tokens": 407, "eval_tokens": 1, "server_latency_ms": 208.0321499997808, "prompt_eval_ms": null, "total_ms": 208.05832999758422, "done_reason": null, "id": "c-einstein-01-dessert-03", "uid": "c022", "task": "kc", "label": "einstein", "text_sha256": "4fae45c9eb873c9e765d8f804b3b5dbb19d1f7b44fcb0911b23c37f40678fd6e", "request_sha256": "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "compiled_tokens": 407, "compiled_prompt_sha256": "7f48d90264fa637fb31af7ae8698c998de26360d2468e492bee50203c0985399", "compiled_ids_sha256": "d0e8df03971a580df0512cfa9b64c5331720d7183da1c0497ce954a5948dbfa7", "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": "2026-09-28T15:32:43Z"} {"mode": "kev-choice", "prompt_sha256": "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "n_options": 6, "kev_response": {"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": 171.96961100125918, "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": 0.1685282879989245, "probabilities_f64_minus_runtime_max_abs": 8.404685836183035e-09}}}, "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": 0.17199459100083914, "prompt_tokens": 299, "eval_tokens": 1, "server_latency_ms": 171.96961100125918, "prompt_eval_ms": null, "total_ms": 171.99459100083914, "done_reason": null, "id": "c-einstein-01-main-01", "uid": "c023", "task": "kc", "label": "einstein", "text_sha256": "07aba25149a0c8da9b647404800cdd24da603dcee8ee4d3dea732f27bcb53bf7", "request_sha256": "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "compiled_tokens": 299, "compiled_prompt_sha256": "398a30a8df523f7677bfca21e4476fae04e1c93179ccaa5354f79989bf208257", "compiled_ids_sha256": "4e8ff629b9c571ecae098a26e5b2d5c09709bf9dbea7851caa35fd43fed66dce", "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": "2026-09-28T15:32:43Z"} {"mode": "kev-choice", "prompt_sha256": "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "n_options": 6, "kev_response": {"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": 182.64648799959105, "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": 0.17858467999758432, "probabilities_f64_minus_runtime_max_abs": 3.111916424547445e-08}}}, "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": 0.1826728779997211, "prompt_tokens": 346, "eval_tokens": 1, "server_latency_ms": 182.64648799959105, "prompt_eval_ms": null, "total_ms": 182.6728779997211, "done_reason": null, "id": "c-einstein-01-main-01", "uid": "c024", "task": "kc", "label": "einstein", "text_sha256": "4f422e6774b951bd581ef4c7876d7e36747ea40e939108fd5e42273d547984fa", "request_sha256": "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "compiled_tokens": 346, "compiled_prompt_sha256": "31ad7358200e1676880e3e7daab4375eb6f0352b2bc77278e208e561783c83f2", "compiled_ids_sha256": "c9faf3bf2ed711f2c81c34c713922081a5f5829813476075c406186f3052df81", "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": "2026-09-28T15:32:43Z"} {"mode": "kev-choice", "prompt_sha256": "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "n_options": 6, "kev_response": {"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": 191.08956200216198, "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": 0.18675799599805032, "probabilities_f64_minus_runtime_max_abs": 3.776358198370744e-08}}}, "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": 0.1911160120034765, "prompt_tokens": 384, "eval_tokens": 1, "server_latency_ms": 191.08956200216198, "prompt_eval_ms": null, "total_ms": 191.1160120034765, "done_reason": null, "id": "c-einstein-01-main-02", "uid": "c025", "task": "kc", "label": "einstein", "text_sha256": "17705c749c69c5c0abdfcac80343c256023b8e361313ef21348a7d1a2fa7ce0a", "request_sha256": "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "compiled_tokens": 384, "compiled_prompt_sha256": "8bc0000acb4ce0e5f74190578a430fa97ca4b3961d56af5f977ae0c41db74aee", "compiled_ids_sha256": "96c01d22640684d10ebbdf561b48b5676cfc3b4bf58f2c5ccd9550d3ab256a3e", "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": "2026-09-28T15:32:43Z"} {"mode": "kev-choice", "prompt_sha256": "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "n_options": 6, "kev_response": {"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": 182.02118299814174, "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": 0.17806432300130837, "probabilities_f64_minus_runtime_max_abs": 9.530233513466158e-08}}}, "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": 0.18204768299983698, "prompt_tokens": 339, "eval_tokens": 1, "server_latency_ms": 182.02118299814174, "prompt_eval_ms": null, "total_ms": 182.04768299983698, "done_reason": null, "id": "c-einstein-01-main-03", "uid": "c026", "task": "kc", "label": "einstein", "text_sha256": "4f6f62821fd6f35d26a1a816268097bf2622a617d550b39a78052b55c661bfc9", "request_sha256": "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "compiled_tokens": 339, "compiled_prompt_sha256": "babae6f4edd94f8ec21e9cf611f98b3fccee444a918dac8ccf6454902374a632", "compiled_ids_sha256": "6866bdd0a49c4b3aa6a808c4deba4a1c9165cab5449d0a129281ece86c4bd505", "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": "2026-09-28T15:32:44Z"} {"mode": "kev-choice", "prompt_sha256": "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "n_options": 6, "kev_response": {"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": 182.80263700216892, "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": 0.17886410699793487, "probabilities_f64_minus_runtime_max_abs": 6.042406330841743e-08}}}, "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": 0.1828280170011567, "prompt_tokens": 345, "eval_tokens": 1, "server_latency_ms": 182.80263700216892, "prompt_eval_ms": null, "total_ms": 182.8280170011567, "done_reason": null, "id": "c-einstein-01-main-03", "uid": "c027", "task": "kc", "label": "einstein", "text_sha256": "8ffb6e325b770fdaeee2ad284e09c015327cc5a633bf8c8c97a5cd3a7b6eea00", "request_sha256": "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "compiled_tokens": 345, "compiled_prompt_sha256": "5b0777e33fab5b56af64797557c932888e6e56a8386aa298f23774efd807c843", "compiled_ids_sha256": "526e5ba6afca227b1070971298c7ab7274443e599e3f4513ee13c7dcebd97fa0", "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": "2026-09-28T15:32:44Z"} {"mode": "kev-choice", "prompt_sha256": "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "n_options": 6, "kev_response": {"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": 188.7774110000464, "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": 0.1848058409996156, "probabilities_f64_minus_runtime_max_abs": 8.697155218229113e-09}}}, "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": 0.18880255999829387, "prompt_tokens": 358, "eval_tokens": 1, "server_latency_ms": 188.7774110000464, "prompt_eval_ms": null, "total_ms": 188.80255999829387, "done_reason": null, "id": "c-einstein-01-main-03", "uid": "c028", "task": "kc", "label": "einstein", "text_sha256": "1969e8278b9b5763c7105922f3230596533913fb049d2aa5046a08e38ab07387", "request_sha256": "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "compiled_tokens": 358, "compiled_prompt_sha256": "23475b0a8910eef4fdc6c4e53fc0b009b8dddcf3b82d3c5b95467ce8d3a7fa19", "compiled_ids_sha256": "74797ebb675858e67a4e4dc5542e122d2f6918fddd42e509d7955ac569f7cd1a", "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": "2026-09-28T15:32:44Z"} {"mode": "kev-choice", "prompt_sha256": "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "n_options": 6, "kev_response": {"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": 208.6239359996398, "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": 0.2040002520006965, "probabilities_f64_minus_runtime_max_abs": 1.4940270798291522e-08}}}, "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": 0.20865157500156783, "prompt_tokens": 402, "eval_tokens": 1, "server_latency_ms": 208.6239359996398, "prompt_eval_ms": null, "total_ms": 208.65157500156783, "done_reason": null, "id": "c-einstein-01-main-03", "uid": "c029", "task": "kc", "label": "einstein", "text_sha256": "09fcaf85056d117752e4cb783fd351b35d15712f09db521a1956d6e0f9c7476e", "request_sha256": "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "compiled_tokens": 402, "compiled_prompt_sha256": "1f30f52d13f815504e7bf1165146b0c24247b90d1788ddb8d1a5b6473d203fe6", "compiled_ids_sha256": "8e661bcb0784f987ceba2e82943447a2e43cffd2e0348e8aba9d9b56cec2f1d5", "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": "2026-09-28T15:32:44Z"} {"mode": "kev-choice", "prompt_sha256": "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "n_options": 6, "kev_response": {"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": 172.99164300129632, "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": 0.16932835099942167, "probabilities_f64_minus_runtime_max_abs": 1.1106184349962689e-07}}}, "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": 0.17301556299935328, "prompt_tokens": 305, "eval_tokens": 1, "server_latency_ms": 172.99164300129632, "prompt_eval_ms": null, "total_ms": 173.01556299935328, "done_reason": null, "id": "c-einstein-01-starter-01", "uid": "c030", "task": "kc", "label": "einstein", "text_sha256": "08e02bf156cc40d29eb210f32660351ca76ed14d072452ab76abe5df36f30cf1", "request_sha256": "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "compiled_tokens": 305, "compiled_prompt_sha256": "53cefe6d23d0ac7f3c46766dc20453dc7cb5499f889b97160ea8df183525bf2a", "compiled_ids_sha256": "f3e5ac56a91af44389f871a8caca89d7e9b0d855b71d7a5f4fe583f155b126f2", "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": "2026-09-28T15:32:44Z"} {"mode": "kev-choice", "prompt_sha256": "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "n_options": 6, "kev_response": {"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": 168.67788699892117, "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": 0.16545711199796642, "probabilities_f64_minus_runtime_max_abs": 4.2042173409306827e-08}}}, "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": 0.1687028870001086, "prompt_tokens": 267, "eval_tokens": 1, "server_latency_ms": 168.67788699892117, "prompt_eval_ms": null, "total_ms": 168.7028870001086, "done_reason": null, "id": "c-einstein-01-starter-02", "uid": "c031", "task": "kc", "label": "einstein", "text_sha256": "a36451d05d663fe9e9be29012d01cc938a3f5f796503647444fa24af9f43c037", "request_sha256": "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "compiled_tokens": 267, "compiled_prompt_sha256": "2c70462d4e85f4e64653498d9e754bf182bcbee31c8f1c5ece9f76079ffaacbf", "compiled_ids_sha256": "fc5d36802beb2fda14c96e74d28c496080c202d461eda3e170bc4d68d9dc027a", "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": "2026-09-28T15:32:45Z"} {"mode": "kev-choice", "prompt_sha256": "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "n_options": 6, "kev_response": {"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": 171.028838998609, "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": 0.16762642499816138, "probabilities_f64_minus_runtime_max_abs": 3.898033895666231e-08}}}, "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": 0.17105195799740613, "prompt_tokens": 294, "eval_tokens": 1, "server_latency_ms": 171.028838998609, "prompt_eval_ms": null, "total_ms": 171.05195799740613, "done_reason": null, "id": "c-einstein-01-starter-02", "uid": "c032", "task": "kc", "label": "einstein", "text_sha256": "514fdcb80f9e03597268ab55271c5a422f6c2fad64c0388d69b1a9fc12a70238", "request_sha256": "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "compiled_tokens": 294, "compiled_prompt_sha256": "370a32f6cf6865231ae43ab9a2c0130f1bd393f6ff9c93f7322c5eec6f69a7f6", "compiled_ids_sha256": "45bd3725a496f6775e0e3d110492d7c95aeb7acbd0c5e55b8758c122228a0ab4", "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": "2026-09-28T15:32:45Z"} {"mode": "kev-choice", "prompt_sha256": "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "n_options": 6, "kev_response": {"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": 206.31031400262145, "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": 0.20207559700065758, "probabilities_f64_minus_runtime_max_abs": 9.327589811647385e-09}}}, "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": 0.20633702399936737, "prompt_tokens": 387, "eval_tokens": 1, "server_latency_ms": 206.31031400262145, "prompt_eval_ms": null, "total_ms": 206.33702399936737, "done_reason": null, "id": "c-einstein-01-starter-03", "uid": "c033", "task": "kc", "label": "einstein", "text_sha256": "b6f364845ed13528ab3c01f57ef05b30b6bf428289727b3932852952d458fa8d", "request_sha256": "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "compiled_tokens": 387, "compiled_prompt_sha256": "32dea34ab78631d1128469f41a9a9b72979f369974ba4b1ff7cdc18dca902d44", "compiled_ids_sha256": "4844b40c337845fad28c4f6e260cd19b165cd0fb8800c232d4d2050f7333a62b", "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": "2026-09-28T15:32:45Z"} {"mode": "kev-choice", "prompt_sha256": "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "n_options": 6, "kev_response": {"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": 182.6692379981978, "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": 0.1787800279998919, "probabilities_f64_minus_runtime_max_abs": 9.1600234064515e-09}}}, "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": 0.1826936069992371, "prompt_tokens": 343, "eval_tokens": 1, "server_latency_ms": 182.6692379981978, "prompt_eval_ms": null, "total_ms": 182.6936069992371, "done_reason": null, "id": "c-einstein-01-starter-03", "uid": "c034", "task": "kc", "label": "einstein", "text_sha256": "410e86ba344ca5a2f88341617f6d75ef59d1ee1233f886347c1cef0b365b938c", "request_sha256": "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "compiled_tokens": 343, "compiled_prompt_sha256": "3048ef259ba2c81faa325128892aa67888378d9abaef64bcc7e191dc46c7930d", "compiled_ids_sha256": "dee894996ae3ae3d47634ad4330ad3215bf4766aff9df1233b15480c06ddd321", "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": "2026-09-28T15:32:45Z"} {"mode": "kev-choice", "prompt_sha256": "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "n_options": 6, "kev_response": {"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": 169.72487900056876, "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": 0.16654187300082413, "probabilities_f64_minus_runtime_max_abs": 3.0369563752685735e-08}}}, "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": 0.1697470789986255, "prompt_tokens": 278, "eval_tokens": 1, "server_latency_ms": 169.72487900056876, "prompt_eval_ms": null, "total_ms": 169.7470789986255, "done_reason": null, "id": "c-einstein-01-starter-03", "uid": "c035", "task": "kc", "label": "einstein", "text_sha256": "67ab5560f37aea121e4812ee7208fe88b05d544de1313ee421b13d1f809a395d", "request_sha256": "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "compiled_tokens": 278, "compiled_prompt_sha256": "8f4e67e82afe236dfe30a2d5eeec24f6f2778b73186fa1f61532a001b7adee48", "compiled_ids_sha256": "bc3a2e6b1255920114fcfe262b849bf8a7b90aa700f811ca5a13e3dea051bd85", "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": "2026-09-28T15:32:45Z"} {"mode": "kev-choice", "prompt_sha256": "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "n_options": 6, "kev_response": {"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": 207.333815000311, "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": 0.20288832100050058, "probabilities_f64_minus_runtime_max_abs": 2.45611590787842e-08}}}, "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": 0.2073592449996795, "prompt_tokens": 395, "eval_tokens": 1, "server_latency_ms": 207.333815000311, "prompt_eval_ms": null, "total_ms": 207.3592449996795, "done_reason": null, "id": "c-hypatia-01-dessert-01", "uid": "c036", "task": "kc", "label": "hypatia", "text_sha256": "a5de183d7f21379964de4518ebf80f3053096a36e1dd875744e7d22e87edc7c1", "request_sha256": "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "compiled_tokens": 395, "compiled_prompt_sha256": "4231600771fa3f6d01b11d9780088dab4c81fa8caa77568d942fc82a7a6b4f8c", "compiled_ids_sha256": "7ef9a0540d1930f31f588689d49c705e105a918541d1c90201320219a03b21c8", "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": "2026-09-28T15:32:46Z"} {"mode": "kev-choice", "prompt_sha256": "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "n_options": 6, "kev_response": {"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": 180.62990399994305, "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": 0.17699619199993322, "probabilities_f64_minus_runtime_max_abs": 1.554243156398627e-08}}}, "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": 0.18066051399728167, "prompt_tokens": 323, "eval_tokens": 1, "server_latency_ms": 180.62990399994305, "prompt_eval_ms": null, "total_ms": 180.66051399728167, "done_reason": null, "id": "c-hypatia-01-dessert-02", "uid": "c037", "task": "kc", "label": "hypatia", "text_sha256": "01a0fda2d7f9187d9c9d4637119020d2e87fb932554aecc4f25d5e1049099cfd", "request_sha256": "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "compiled_tokens": 323, "compiled_prompt_sha256": "64cf4ddb66e4162fd4413e7c5eecfae78ed5607c7644b898fafd96f028bf8081", "compiled_ids_sha256": "2edffcb25f65010db4bf8ec46a80e7f11423c9168ec99e3b3ac49296338f2c48", "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": "2026-09-28T15:32:46Z"} {"mode": "kev-choice", "prompt_sha256": "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "n_options": 6, "kev_response": {"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": 169.47916000208352, "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": 0.16634083499957342, "probabilities_f64_minus_runtime_max_abs": 3.017469207833301e-08}}}, "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": 0.16950218000056338, "prompt_tokens": 274, "eval_tokens": 1, "server_latency_ms": 169.47916000208352, "prompt_eval_ms": null, "total_ms": 169.50218000056338, "done_reason": null, "id": "c-hypatia-01-dessert-02", "uid": "c038", "task": "kc", "label": "hypatia", "text_sha256": "d19c1175d3216a8a0bfd4d0953337a3682e2eacf13a33a217ea2c81dacc6ba81", "request_sha256": "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "compiled_tokens": 274, "compiled_prompt_sha256": "b813667369f4e70abaa4bbd36b3b8b690e78c38b9ec85f4f3d5388a491e9ce79", "compiled_ids_sha256": "1356d08876a1c4cf2b6410ee614660ec7bf937bb468b506992fcb7bf08d22f64", "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": "2026-09-28T15:32:46Z"} {"mode": "kev-choice", "prompt_sha256": "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "n_options": 6, "kev_response": {"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": 171.53180499735754, "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": 0.16811097100071493, "probabilities_f64_minus_runtime_max_abs": 1.7241397975986672e-08}}}, "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": 0.17155520499727572, "prompt_tokens": 298, "eval_tokens": 1, "server_latency_ms": 171.53180499735754, "prompt_eval_ms": null, "total_ms": 171.55520499727572, "done_reason": null, "id": "c-hypatia-01-dessert-03", "uid": "c039", "task": "kc", "label": "hypatia", "text_sha256": "5f26c48e950fedd038bfdac0246a243725b4544c160a6d2b3bdc1bc678cd5d07", "request_sha256": "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "compiled_tokens": 298, "compiled_prompt_sha256": "21829cbfc35a4be0d7b256334b6e53727cf63ff22411b3b96a97b55be3b983e1", "compiled_ids_sha256": "90ee18b88d0e89b3999ace8ac624b7a7824208a64fcb7a98548c6bbcf7208086", "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": "2026-09-28T15:32:46Z"} {"mode": "kev-choice", "prompt_sha256": "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "n_options": 6, "kev_response": {"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": 190.08795999980066, "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": 0.18604383199999575, "probabilities_f64_minus_runtime_max_abs": 1.7455427769874632e-08}}}, "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": 0.19011237000086112, "prompt_tokens": 368, "eval_tokens": 1, "server_latency_ms": 190.08795999980066, "prompt_eval_ms": null, "total_ms": 190.11237000086112, "done_reason": null, "id": "c-hypatia-01-main-01", "uid": "c040", "task": "kc", "label": "hypatia", "text_sha256": "6d5bf83d9dbfa6c8fa54c2ac6f17730078b8b12a6895818e9d5bae99813c9910", "request_sha256": "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "compiled_tokens": 368, "compiled_prompt_sha256": "a85f301fbcbe5359d5137e8ce9e30a12f45037248d0741358fb7ff25b1b3eafd", "compiled_ids_sha256": "c6dbc71f4fcb0e3fc82bd7695103d89d45b3a9f018a7f14f1ed0ba8bed8dbf04", "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": "2026-09-28T15:32:46Z"} {"mode": "kev-choice", "prompt_sha256": "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "n_options": 6, "kev_response": {"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": 171.65122400183463, "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": 0.16825937000248814, "probabilities_f64_minus_runtime_max_abs": 1.8731469264565703e-08}}}, "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": 0.17167454300215468, "prompt_tokens": 297, "eval_tokens": 1, "server_latency_ms": 171.65122400183463, "prompt_eval_ms": null, "total_ms": 171.67454300215468, "done_reason": null, "id": "c-hypatia-01-main-01", "uid": "c041", "task": "kc", "label": "hypatia", "text_sha256": "f1b2cfbff4f5841853cb9507da2768ff094aeeca9b559b51245ee33878dffeb2", "request_sha256": "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "compiled_tokens": 297, "compiled_prompt_sha256": "937427227f5f99118cb41f68bee040c493d9bec3da24346374497e87fd4b0c16", "compiled_ids_sha256": "f2a92c82350a676ec89ba5633141c479e87276520dc9842cd8d2d1356c426518", "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": "2026-09-28T15:32:46Z"} {"mode": "kev-choice", "prompt_sha256": "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "n_options": 6, "kev_response": {"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": 206.3929329997336, "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": 0.2020540270023048, "probabilities_f64_minus_runtime_max_abs": 2.473366109145836e-08}}}, "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": 0.206419073001598, "prompt_tokens": 388, "eval_tokens": 1, "server_latency_ms": 206.3929329997336, "prompt_eval_ms": null, "total_ms": 206.419073001598, "done_reason": null, "id": "c-hypatia-01-main-01", "uid": "c042", "task": "kc", "label": "hypatia", "text_sha256": "e252c70e869af17f3ef75ee21e1cac261c8c996f84df756e6abdf0a420baf24b", "request_sha256": "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "compiled_tokens": 388, "compiled_prompt_sha256": "9e620d83c4565136473c23f396206defac1ffa16bd69dabc4a8fb76aa81893ae", "compiled_ids_sha256": "5cdc4054cab42c152b7033db49be91482544fff96b21837effc480511601a69b", "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": "2026-09-28T15:32:47Z"} {"mode": "kev-choice", "prompt_sha256": "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "n_options": 6, "kev_response": {"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": 182.5884079989919, "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": 0.17872135899960995, "probabilities_f64_minus_runtime_max_abs": 1.196630311506297e-08}}}, "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": 0.18261331800022162, "prompt_tokens": 339, "eval_tokens": 1, "server_latency_ms": 182.5884079989919, "prompt_eval_ms": null, "total_ms": 182.61331800022162, "done_reason": null, "id": "c-hypatia-01-main-01", "uid": "c043", "task": "kc", "label": "hypatia", "text_sha256": "c9971b306d10c18655f624323a0657456d02d8b9f6f8c8a390c53a0144a20353", "request_sha256": "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "compiled_tokens": 339, "compiled_prompt_sha256": "96b3447c5f6b1e9a91ad93d8847b48d6895167515fc8658b556b546cd722af75", "compiled_ids_sha256": "a5315fe99a572917cd0eab60dbea5014cc09871c3719dbb84d88bd7901126ebf", "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": "2026-09-28T15:32:47Z"} {"mode": "kev-choice", "prompt_sha256": "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "n_options": 6, "kev_response": {"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": 169.7273089994269, "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": 0.16652775399779784, "probabilities_f64_minus_runtime_max_abs": 2.561325496142075e-08}}}, "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": 0.16975023800114286, "prompt_tokens": 275, "eval_tokens": 1, "server_latency_ms": 169.7273089994269, "prompt_eval_ms": null, "total_ms": 169.75023800114286, "done_reason": null, "id": "c-hypatia-01-main-01", "uid": "c044", "task": "kc", "label": "hypatia", "text_sha256": "241d6cdc11e0749d2f38d145fbd4bda015a93b0c01b2e0692f7d374cdd1fc0a9", "request_sha256": "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "compiled_tokens": 275, "compiled_prompt_sha256": "12910ec5c352e9fa1330272cb09c7a1e3ca9ef566bd28df789137a8794a156bc", "compiled_ids_sha256": "bf822f6c8b673ada113561218d839770895abe054bccedd183916316a0e84576", "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": "2026-09-28T15:32:47Z"} {"mode": "kev-choice", "prompt_sha256": "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "n_options": 6, "kev_response": {"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": 182.65155799963395, "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": 0.1787129289987206, "probabilities_f64_minus_runtime_max_abs": 1.7540869451382335e-08}}}, "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": 0.18267501700029243, "prompt_tokens": 349, "eval_tokens": 1, "server_latency_ms": 182.65155799963395, "prompt_eval_ms": null, "total_ms": 182.67501700029243, "done_reason": null, "id": "c-hypatia-01-main-02", "uid": "c045", "task": "kc", "label": "hypatia", "text_sha256": "990f4efb9ecf643ec7583c478c70e24062ee44b2b11fc92c39414c900af4f50c", "request_sha256": "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "compiled_tokens": 349, "compiled_prompt_sha256": "53ea90fe5d10904b59900dd3288e3f69d354a8b53be8832d2ed7310f00a05a78", "compiled_ids_sha256": "fc14aba23487d6da377ebf5043577838bb63cffc65b91089be20b810df6d2bad", "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": "2026-09-28T15:32:47Z"} {"mode": "kev-choice", "prompt_sha256": "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "n_options": 6, "kev_response": {"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": 190.05072099753306, "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": 0.18604367100124364, "probabilities_f64_minus_runtime_max_abs": 2.120121542215081e-08}}}, "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": 0.1900758610026969, "prompt_tokens": 364, "eval_tokens": 1, "server_latency_ms": 190.05072099753306, "prompt_eval_ms": null, "total_ms": 190.0758610026969, "done_reason": null, "id": "c-hypatia-01-main-03", "uid": "c046", "task": "kc", "label": "hypatia", "text_sha256": "2f798896840b41edb036a413b36f53af135fdff331b30e07736bc0e6c8953bd7", "request_sha256": "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "compiled_tokens": 364, "compiled_prompt_sha256": "981f394b92cdbfe83d1629e86ff1043844d3ddec1aa960eee15412baa3b51c00", "compiled_ids_sha256": "e1be0ad9a565296219335b76cdbbdeef9cf60282395d86cca26940dc6907053b", "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": "2026-09-28T15:32:47Z"} {"mode": "kev-choice", "prompt_sha256": "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "n_options": 6, "kev_response": {"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": 180.83524299800047, "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": 0.17714882099971874, "probabilities_f64_minus_runtime_max_abs": 4.094242278540605e-08}}}, "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": 0.18086108099669218, "prompt_tokens": 329, "eval_tokens": 1, "server_latency_ms": 180.83524299800047, "prompt_eval_ms": null, "total_ms": 180.86108099669218, "done_reason": null, "id": "c-hypatia-01-main-03", "uid": "c047", "task": "kc", "label": "hypatia", "text_sha256": "76cf8cc11125b2570e9c8e373627b9b8bce42a3ed2f50594fad3be025b99d1a8", "request_sha256": "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "compiled_tokens": 329, "compiled_prompt_sha256": "5e16fbf7126da6744631cdedb3c9e498a2b9052c185376919d00b1f279b12f9c", "compiled_ids_sha256": "07c699060a078fa6ae02cca28988321df21cdd9ed54996d09648ee655e350dce", "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": "2026-09-28T15:32:48Z"} {"mode": "kev-choice", "prompt_sha256": "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "n_options": 6, "kev_response": {"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": 208.60706600069534, "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": 0.20399198200175306, "probabilities_f64_minus_runtime_max_abs": 6.771495131285832e-08}}}, "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": 0.20863588499923935, "prompt_tokens": 407, "eval_tokens": 1, "server_latency_ms": 208.60706600069534, "prompt_eval_ms": null, "total_ms": 208.63588499923935, "done_reason": null, "id": "c-hypatia-01-main-03", "uid": "c048", "task": "kc", "label": "hypatia", "text_sha256": "50041391af6fba1f4bd767736233b10e810a2891780598b23f019220d35ebab1", "request_sha256": "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "compiled_tokens": 407, "compiled_prompt_sha256": "1758a9f1601e86873bbd3a7f9387586fd8779e1252ba78aa05c488753ae18af9", "compiled_ids_sha256": "c2e4ce3277e1b394388c8fbe6816aea6f422bb1f7b5f4acc516099b73569fad1", "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": "2026-09-28T15:32:48Z"} {"mode": "kev-choice", "prompt_sha256": "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "n_options": 6, "kev_response": {"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": 169.79499800072517, "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": 0.16658546300095622, "probabilities_f64_minus_runtime_max_abs": 4.581853807472669e-08}}}, "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": 0.1698185579989513, "prompt_tokens": 272, "eval_tokens": 1, "server_latency_ms": 169.79499800072517, "prompt_eval_ms": null, "total_ms": 169.8185579989513, "done_reason": null, "id": "c-hypatia-01-starter-01", "uid": "c049", "task": "kc", "label": "hypatia", "text_sha256": "b3cbd4a18274f64a6bde319e80f25054e12322fb09acc0c3d522792b7b49ede8", "request_sha256": "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "compiled_tokens": 272, "compiled_prompt_sha256": "86edd11982c21ff733510b65fb8061cf99f203a5783f0d267f8b7922baef66d7", "compiled_ids_sha256": "0632451188e0cd8274e38703ba69b1d81b731677e7550055093537363c0d1780", "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": "2026-09-28T15:32:48Z"} {"mode": "kev-choice", "prompt_sha256": "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "n_options": 6, "kev_response": {"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": 180.24863699974958, "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": 0.17661973500071326, "probabilities_f64_minus_runtime_max_abs": 2.920105179882171e-08}}}, "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": 0.18027165700186742, "prompt_tokens": 326, "eval_tokens": 1, "server_latency_ms": 180.24863699974958, "prompt_eval_ms": null, "total_ms": 180.27165700186742, "done_reason": null, "id": "c-hypatia-01-starter-01", "uid": "c050", "task": "kc", "label": "hypatia", "text_sha256": "6588c54ef23e640e7f57252c699affd5dd6b2af55b7d2263da46b57ca15b64fd", "request_sha256": "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "compiled_tokens": 326, "compiled_prompt_sha256": "ecbc5708033f479cd2caa368a60bec322a499a7331fb8a6eacdc2dac8b96a819", "compiled_ids_sha256": "3213fbf2a47076131e60a1856d79082d20d04b59cb63f46819df660c2b4096ff", "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": "2026-09-28T15:32:48Z"} {"mode": "kev-choice", "prompt_sha256": "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "n_options": 6, "kev_response": {"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": 172.1792300013476, "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": 0.1688454549985181, "probabilities_f64_minus_runtime_max_abs": 2.27861085555503e-08}}}, "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": 0.17220183900280972, "prompt_tokens": 303, "eval_tokens": 1, "server_latency_ms": 172.1792300013476, "prompt_eval_ms": null, "total_ms": 172.20183900280972, "done_reason": null, "id": "c-hypatia-01-starter-02", "uid": "c051", "task": "kc", "label": "hypatia", "text_sha256": "584240bc99dcaa76ab12a358683ec6fe3c209a90fc88a943d6e23be4b03fd7fe", "request_sha256": "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "compiled_tokens": 303, "compiled_prompt_sha256": "d3cc18c537cd2524953ff31d3b343b481a694d8832a6516cf95f0af6ac3a93fc", "compiled_ids_sha256": "4176637f3b7f7e0c1eaf039089f3e571cafaabcf2dff27018fe9ae1943bf4ef4", "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": "2026-09-28T15:32:48Z"} {"mode": "kev-choice", "prompt_sha256": "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "n_options": 6, "kev_response": {"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": 169.1098230003263, "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": 0.1659668079992116, "probabilities_f64_minus_runtime_max_abs": 2.1162001595520508e-08}}}, "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": 0.16913180299889063, "prompt_tokens": 273, "eval_tokens": 1, "server_latency_ms": 169.1098230003263, "prompt_eval_ms": null, "total_ms": 169.13180299889063, "done_reason": null, "id": "c-hypatia-01-starter-02", "uid": "c052", "task": "kc", "label": "hypatia", "text_sha256": "e91537d8f3532b1cdb47129a57ad062155f8ced9d488616dd4276ec4164938fb", "request_sha256": "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "compiled_tokens": 273, "compiled_prompt_sha256": "25078b32820a0c397cc2958b1cb9a27364e725d270a1e669793c995eab4ad072", "compiled_ids_sha256": "9403e074d597096fdde8ed17e223082b95d8f4cb2f6a7e59d08436d83892a5ca", "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": "2026-09-28T15:32:48Z"} {"mode": "kev-choice", "prompt_sha256": "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "n_options": 6, "kev_response": {"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": 172.64797600000747, "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": 0.1691701430027024, "probabilities_f64_minus_runtime_max_abs": 3.4154779338102514e-08}}}, "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": 0.1726701259976835, "prompt_tokens": 309, "eval_tokens": 1, "server_latency_ms": 172.64797600000747, "prompt_eval_ms": null, "total_ms": 172.6701259976835, "done_reason": null, "id": "c-hypatia-01-starter-02", "uid": "c053", "task": "kc", "label": "hypatia", "text_sha256": "8d7b0004f9f8bdcc93faf53126599249fb43b754daa1ab89d93c77cb67b3310d", "request_sha256": "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "compiled_tokens": 309, "compiled_prompt_sha256": "ae3c350b97daa5c6854c06d3ed1c5bfc8800493c92943b74514bed4d49857d5a", "compiled_ids_sha256": "55a7c5fbd189b1b2d5df695ec9fc001b672fd7fe50d6a7ee300faac082ac3ee1", "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": "2026-09-28T15:32:49Z"} {"mode": "kev-choice", "prompt_sha256": "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "n_options": 6, "kev_response": {"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": 173.36349999823142, "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": 0.1696866090023832, "probabilities_f64_minus_runtime_max_abs": 2.491715117569626e-08}}}, "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": 0.17338638000001083, "prompt_tokens": 311, "eval_tokens": 1, "server_latency_ms": 173.36349999823142, "prompt_eval_ms": null, "total_ms": 173.38638000001083, "done_reason": null, "id": "c-ibn_sina-01-dessert-01", "uid": "c054", "task": "kc", "label": "ibn_sina", "text_sha256": "b299c905839b0a3ea80642f3563ed7da112315705aac5520df3856a5ae1de6c9", "request_sha256": "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "compiled_tokens": 311, "compiled_prompt_sha256": "066735836834e07c7890804951ba5a2894b840544a1b10249ec2970aea3a1c62", "compiled_ids_sha256": "8b485c39edfb00d587270bcc96d96b07d12859a38e102a757f143839cbda8eb7", "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": "2026-09-28T15:32:49Z"} {"mode": "kev-choice", "prompt_sha256": "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "n_options": 6, "kev_response": {"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": 180.01721800101222, "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": 0.17631190700194566, "probabilities_f64_minus_runtime_max_abs": 2.65361272955289e-08}}}, "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": 0.18004062800173415, "prompt_tokens": 323, "eval_tokens": 1, "server_latency_ms": 180.01721800101222, "prompt_eval_ms": null, "total_ms": 180.04062800173415, "done_reason": null, "id": "c-ibn_sina-01-dessert-02", "uid": "c055", "task": "kc", "label": "ibn_sina", "text_sha256": "08a2d58741f0e2b5e18564f563f0c14b6a31eedffbd7d8bd3650b24b79ed9e26", "request_sha256": "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "compiled_tokens": 323, "compiled_prompt_sha256": "aa541bf54e4da10cb7233c63ec2579c3ba492ff5efbefd219bec4731f7f5d42b", "compiled_ids_sha256": "6d3e88b7607381a1857be3378d363a3a49ad70080bd24153eee4a6f79fe0c29c", "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": "2026-09-28T15:32:49Z"} {"mode": "kev-choice", "prompt_sha256": "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "n_options": 6, "kev_response": {"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": 182.45157999990624, "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": 0.1786593590004486, "probabilities_f64_minus_runtime_max_abs": 2.5152750271484337e-08}}}, "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": 0.1824753300024895, "prompt_tokens": 342, "eval_tokens": 1, "server_latency_ms": 182.45157999990624, "prompt_eval_ms": null, "total_ms": 182.4753300024895, "done_reason": null, "id": "c-ibn_sina-01-dessert-02", "uid": "c056", "task": "kc", "label": "ibn_sina", "text_sha256": "662296bb66a18a37300fd190cf48c9853e307386b2ca49c84cb14a90465ba367", "request_sha256": "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "compiled_tokens": 342, "compiled_prompt_sha256": "725690d1f22c93d0340b285452fdb898cbeffbf60adbcee2a6cf9213e82abbc0", "compiled_ids_sha256": "94b95ec2aa38d8983aa6f7c9626041fbc139ec74d98175507b3c7f676e48c4ab", "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": "2026-09-28T15:32:49Z"} {"mode": "kev-choice", "prompt_sha256": "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "n_options": 6, "kev_response": {"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": 182.29029000212904, "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": 0.17854616999829886, "probabilities_f64_minus_runtime_max_abs": 9.904772779734117e-09}}}, "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": 0.18231410999942455, "prompt_tokens": 339, "eval_tokens": 1, "server_latency_ms": 182.29029000212904, "prompt_eval_ms": null, "total_ms": 182.31410999942455, "done_reason": null, "id": "c-ibn_sina-01-dessert-02", "uid": "c057", "task": "kc", "label": "ibn_sina", "text_sha256": "d8e0cb07e0eaf48824f57ebe5eb906dcef9a8f1e4b6f099dff399409791584b6", "request_sha256": "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "compiled_tokens": 339, "compiled_prompt_sha256": "319fd600c27e869175120d52429454dc20b5548f94e0e370905b96ceb450ebfc", "compiled_ids_sha256": "7cd4b48291d7d001e70cad1045450e5ebc69a94b89e94cfc8233e6b63d6d448e", "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": "2026-09-28T15:32:49Z"} {"mode": "kev-choice", "prompt_sha256": "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "n_options": 6, "kev_response": {"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": 183.18492399703246, "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": 0.17930067400084226, "probabilities_f64_minus_runtime_max_abs": 6.539001840000225e-09}}}, "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": 0.18320910399779677, "prompt_tokens": 351, "eval_tokens": 1, "server_latency_ms": 183.18492399703246, "prompt_eval_ms": null, "total_ms": 183.20910399779677, "done_reason": null, "id": "c-ibn_sina-01-dessert-03", "uid": "c058", "task": "kc", "label": "ibn_sina", "text_sha256": "f4471ae63e6d6bccbd5c261f4532ca1a299e3f457f483d8761a8918386e82bc1", "request_sha256": "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "compiled_tokens": 351, "compiled_prompt_sha256": "31e37b8f5f725ef736f2d61439d483950b0c0100444ccf92ee30ba189e9af0ef", "compiled_ids_sha256": "a7851da9192bd84a58d364630f143cf2398c247083217825c24b2c65fd8f0f9e", "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": "2026-09-28T15:32:49Z"} {"mode": "kev-choice", "prompt_sha256": "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "n_options": 6, "kev_response": {"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": 207.24698699996225, "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": 0.2029252499996801, "probabilities_f64_minus_runtime_max_abs": 8.229804615567105e-09}}}, "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": 0.20727171700127656, "prompt_tokens": 393, "eval_tokens": 1, "server_latency_ms": 207.24698699996225, "prompt_eval_ms": null, "total_ms": 207.27171700127656, "done_reason": null, "id": "c-ibn_sina-01-dessert-03", "uid": "c059", "task": "kc", "label": "ibn_sina", "text_sha256": "bd105be757c9e850ab01e91c392d5b3ffcc1a669e30151f6297ad7d4a9f43e50", "request_sha256": "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "compiled_tokens": 393, "compiled_prompt_sha256": "ba26f55ee8938b9cb0c3d4aabab1559ed19fcddd55502d02f38ac6dfdc9310ac", "compiled_ids_sha256": "8f0bad5ac64183cef3d192f03fe419ee35d679774c7b48af85a668fb1670c955", "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": "2026-09-28T15:32:50Z"} {"mode": "kev-choice", "prompt_sha256": "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "n_options": 6, "kev_response": {"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": 174.2948230021284, "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": 0.17052028300167876, "probabilities_f64_minus_runtime_max_abs": 1.8511987526226292e-08}}}, "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": 0.17432161299802829, "prompt_tokens": 317, "eval_tokens": 1, "server_latency_ms": 174.2948230021284, "prompt_eval_ms": null, "total_ms": 174.3216129980283, "done_reason": null, "id": "c-ibn_sina-01-main-02", "uid": "c060", "task": "kc", "label": "ibn_sina", "text_sha256": "b55dbe714cc75405acc2e9cc1648dc3ee39c13f22e834fb80f8422ba653843e2", "request_sha256": "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "compiled_tokens": 317, "compiled_prompt_sha256": "ca53ab37565565807844480c303f0ac73c53fed29285230b4137aee8eb216a3b", "compiled_ids_sha256": "46e60b1b41b20e85a4c648d7194e57c8c2da9d2e1ec5c2d9640a9a1b77e6ad6e", "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": "2026-09-28T15:32:50Z"} {"mode": "kev-choice", "prompt_sha256": "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "n_options": 6, "kev_response": {"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": 183.2301340000413, "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": 0.1791927250014851, "probabilities_f64_minus_runtime_max_abs": 5.636584776702591e-08}}}, "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": 0.18325566300336504, "prompt_tokens": 352, "eval_tokens": 1, "server_latency_ms": 183.2301340000413, "prompt_eval_ms": null, "total_ms": 183.25566300336504, "done_reason": null, "id": "c-ibn_sina-01-main-02", "uid": "c061", "task": "kc", "label": "ibn_sina", "text_sha256": "dd2f74b23716b02aa47bf6643e04dff6473ac0c4707db2e339a44e5a0a841efc", "request_sha256": "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "compiled_tokens": 352, "compiled_prompt_sha256": "45e5e8ec908f7291f32e80dc6bdc7bd7e15ac15096efb3d82a2feb0b9cc1ed9e", "compiled_ids_sha256": "8cb23d0f97d0241542c05f0a7bceae8684c28142b6f39ba5f8ef5db574533fac", "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": "2026-09-28T15:32:50Z"} {"mode": "kev-choice", "prompt_sha256": "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "n_options": 6, "kev_response": {"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": 190.13972999891848, "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": 0.18609403099981137, "probabilities_f64_minus_runtime_max_abs": 2.2737410398931956e-08}}}, "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": 0.1901641889999155, "prompt_tokens": 365, "eval_tokens": 1, "server_latency_ms": 190.13972999891848, "prompt_eval_ms": null, "total_ms": 190.1641889999155, "done_reason": null, "id": "c-ibn_sina-01-main-02", "uid": "c062", "task": "kc", "label": "ibn_sina", "text_sha256": "7017c582c8c06a8055ff39050ceb8b9c6ecdfd1950ea6fb82eaa12427352e499", "request_sha256": "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "compiled_tokens": 365, "compiled_prompt_sha256": "2c19043fdf98431a6068bfeda9e535e1c58513f3ad737c128fbdbff9a34cc6d6", "compiled_ids_sha256": "3d9a463164b741f3095477f08090e3b5559a38fc08b00bb779a1cc9e5a019438", "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": "2026-09-28T15:32:50Z"} {"mode": "kev-choice", "prompt_sha256": "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "n_options": 6, "kev_response": {"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": 207.30700599960983, "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": 0.20297327899970696, "probabilities_f64_minus_runtime_max_abs": 2.200523885331762e-08}}}, "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": 0.20733427600134746, "prompt_tokens": 390, "eval_tokens": 1, "server_latency_ms": 207.30700599960983, "prompt_eval_ms": null, "total_ms": 207.33427600134746, "done_reason": null, "id": "c-ibn_sina-01-main-03", "uid": "c063", "task": "kc", "label": "ibn_sina", "text_sha256": "0b11aa8d00f225a2e23695d1879b0b3cfecbb79f138d13ac59b948622fe38cae", "request_sha256": "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "compiled_tokens": 390, "compiled_prompt_sha256": "7b9355b2e2fdcf6d623298737a222d53b12dc9d6c71de4bf927676f9c9b5c339", "compiled_ids_sha256": "45337ce0418c787e52b602dcd036092f44cf3be9bf63859bfa2bb7606fe58827", "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": "2026-09-28T15:32:50Z"} {"mode": "kev-choice", "prompt_sha256": "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "n_options": 6, "kev_response": {"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": 190.29767800020636, "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": 0.1861910000006901, "probabilities_f64_minus_runtime_max_abs": 1.3127908066179472e-08}}}, "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": 0.19032742800118285, "prompt_tokens": 371, "eval_tokens": 1, "server_latency_ms": 190.29767800020636, "prompt_eval_ms": null, "total_ms": 190.32742800118285, "done_reason": null, "id": "c-ibn_sina-01-main-03", "uid": "c064", "task": "kc", "label": "ibn_sina", "text_sha256": "a7d0bd3209b1bd2142727de630b760b15f5f6312081717f4e82e1ee2a0212263", "request_sha256": "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "compiled_tokens": 371, "compiled_prompt_sha256": "0dc80c8adce05b741ce8a4c3f17b9cc1f78bc491ba12ae7dfad74cd0ca49ecbb", "compiled_ids_sha256": "5948b85d5511809552d45e91339724803cfce0ca4a60973ed39e61cc07e354a8", "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": "2026-09-28T15:32:51Z"} {"mode": "kev-choice", "prompt_sha256": "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "n_options": 6, "kev_response": {"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": 206.27553400117904, "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": 0.20182882800145308, "probabilities_f64_minus_runtime_max_abs": 3.6675833370303224e-08}}}, "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": 0.20630158399944776, "prompt_tokens": 395, "eval_tokens": 1, "server_latency_ms": 206.27553400117904, "prompt_eval_ms": null, "total_ms": 206.30158399944776, "done_reason": null, "id": "c-ibn_sina-01-main-03", "uid": "c065", "task": "kc", "label": "ibn_sina", "text_sha256": "3b833e5a5504cbd80f178ad75b526e1f554adfbb0089f158306ea21143c7ae2f", "request_sha256": "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "compiled_tokens": 395, "compiled_prompt_sha256": "667abdaec59327957c3b1ce1c6570cf662d99312a7e39336fd34c6a5b9174833", "compiled_ids_sha256": "5cf614b764b9c0e5e91de93bb9bc44413b47819378f9850fb3926ac12e4ae96b", "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": "2026-09-28T15:32:51Z"} {"mode": "kev-choice", "prompt_sha256": "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "n_options": 6, "kev_response": {"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": 170.5286119977245, "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": 0.16701938999904087, "probabilities_f64_minus_runtime_max_abs": 3.604682752289534e-08}}}, "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": 0.17055105200051912, "prompt_tokens": 293, "eval_tokens": 1, "server_latency_ms": 170.5286119977245, "prompt_eval_ms": null, "total_ms": 170.55105200051912, "done_reason": null, "id": "c-ibn_sina-01-starter-01", "uid": "c066", "task": "kc", "label": "ibn_sina", "text_sha256": "23e51d7b6856050d91d3999166ecb603fc4961be70cc59ae3729087202ef3384", "request_sha256": "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "compiled_tokens": 293, "compiled_prompt_sha256": "0cc652ed158d4456d057ae4ea6f240eebcd0b64c04d9666f8ff066084455a03d", "compiled_ids_sha256": "d87679564b3455f5e970ac7dffd32d0e50699c17f522bec77e89352be1475266", "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": "2026-09-28T15:32:51Z"} {"mode": "kev-choice", "prompt_sha256": "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "n_options": 6, "kev_response": {"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": 172.60528600309044, "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": 0.16903775299942936, "probabilities_f64_minus_runtime_max_abs": 7.863258169593657e-09}}}, "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": 0.17262884599767858, "prompt_tokens": 311, "eval_tokens": 1, "server_latency_ms": 172.60528600309044, "prompt_eval_ms": null, "total_ms": 172.62884599767858, "done_reason": null, "id": "c-ibn_sina-01-starter-02", "uid": "c067", "task": "kc", "label": "ibn_sina", "text_sha256": "29e6791f3ab096493a37cc56ce2519a4c5cb87aafeed65dadcdbcbc34defc75c", "request_sha256": "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "compiled_tokens": 311, "compiled_prompt_sha256": "ecd96e4320ca6e089fea50dd88dd138efc7728b40394e645c167cf30bfe00de9", "compiled_ids_sha256": "ea289ecc97718490e97f10028ee7f3325d7a73f38345e7ea80efb30da1bcf2a9", "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": "2026-09-28T15:32:51Z"} {"mode": "kev-choice", "prompt_sha256": "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "n_options": 6, "kev_response": {"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": 189.53288400007295, "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": 0.1853851570012921, "probabilities_f64_minus_runtime_max_abs": 2.733875414939746e-08}}}, "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": 0.18955774400092196, "prompt_tokens": 371, "eval_tokens": 1, "server_latency_ms": 189.53288400007295, "prompt_eval_ms": null, "total_ms": 189.55774400092196, "done_reason": null, "id": "c-ibn_sina-01-starter-03", "uid": "c068", "task": "kc", "label": "ibn_sina", "text_sha256": "92a317e3024d7f5dbd04f4abe6e6b2dbc9a2b7b76aef5d19ef117adbe1abff1f", "request_sha256": "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "compiled_tokens": 371, "compiled_prompt_sha256": "0ad9f34f9bb88bf2b729ba7edfdf37192f81ae6be3b2bb68a023efff32a6297e", "compiled_ids_sha256": "1fa8144179d5b9344965adbb89b17f9bf2ac697e24f6cd4b3cb376295fba968c", "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": "2026-09-28T15:32:51Z"} {"mode": "kev-choice", "prompt_sha256": "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "n_options": 6, "kev_response": {"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": 182.69789799887803, "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": 0.17874025799756055, "probabilities_f64_minus_runtime_max_abs": 2.8342562741201505e-08}}}, "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": 0.18272462800086942, "prompt_tokens": 355, "eval_tokens": 1, "server_latency_ms": 182.69789799887803, "prompt_eval_ms": null, "total_ms": 182.72462800086942, "done_reason": null, "id": "c-ibn_sina-01-starter-03", "uid": "c069", "task": "kc", "label": "ibn_sina", "text_sha256": "2b38a3ae73d841c463a00f9961952498fec58ca7e4a8afc6ad0d68d3f2da0891", "request_sha256": "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "compiled_tokens": 355, "compiled_prompt_sha256": "02824b55841c1a917666db92f2dab2f28df3699581fc36e4d28cc000f8f4a105", "compiled_ids_sha256": "052585d008f125e9a9c56574e6532af6ea4dffeaa060a8ab415742f7363867e1", "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": "2026-09-28T15:32:52Z"} {"mode": "kev-choice", "prompt_sha256": "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "n_options": 6, "kev_response": {"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": 189.47643499996047, "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": 0.18534163699951023, "probabilities_f64_minus_runtime_max_abs": 1.7496209370193583e-08}}}, "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": 0.18950168400260736, "prompt_tokens": 371, "eval_tokens": 1, "server_latency_ms": 189.47643499996047, "prompt_eval_ms": null, "total_ms": 189.50168400260736, "done_reason": null, "id": "c-ibn_sina-01-starter-03", "uid": "c070", "task": "kc", "label": "ibn_sina", "text_sha256": "e7fbfb77c803a1b49ac58c6a9838ff600043b6d05103c23e581b0222b199442d", "request_sha256": "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "compiled_tokens": 371, "compiled_prompt_sha256": "646ca9742583b77ea42ed581986481a7143ca23eac7d883e38862c3cb4e2dc56", "compiled_ids_sha256": "823c9e43923cf82515b47c19a4962b220c53c5e9d09cfb9b90c5cd78de0ad4e1", "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": "2026-09-28T15:32:52Z"} {"mode": "kev-choice", "prompt_sha256": "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "n_options": 6, "kev_response": {"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": 181.46789700040244, "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": 0.1776282770006219, "probabilities_f64_minus_runtime_max_abs": 3.087528011036511e-08}}}, "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": 0.1814921770019282, "prompt_tokens": 346, "eval_tokens": 1, "server_latency_ms": 181.46789700040244, "prompt_eval_ms": null, "total_ms": 181.4921770019282, "done_reason": null, "id": "c-ibn_sina-01-starter-03", "uid": "c071", "task": "kc", "label": "ibn_sina", "text_sha256": "75fffc308492668b7b9b1c55a6949b5364b630213ba04b50c512ebe89e45e583", "request_sha256": "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "compiled_tokens": 346, "compiled_prompt_sha256": "672de484d211354cd558d5df735a439daa860248a900b0c7bbcb4c54d1e2256c", "compiled_ids_sha256": "1d298de3fe49f6c758af9b8bfaf818d27668637424e77a80f4a96f9bfce9406e", "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": "2026-09-28T15:32:52Z"} {"mode": "kev-choice", "prompt_sha256": "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "n_options": 6, "kev_response": {"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": 207.1998169994913, "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": 0.20258714300143765, "probabilities_f64_minus_runtime_max_abs": 2.6626056137679655e-08}}}, "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": 0.20723028700012947, "prompt_tokens": 404, "eval_tokens": 1, "server_latency_ms": 207.1998169994913, "prompt_eval_ms": null, "total_ms": 207.23028700012947, "done_reason": null, "id": "c-sagan-01-dessert-01", "uid": "c072", "task": "kc", "label": "sagan", "text_sha256": "17954c8ac9ae5c7fedfad2b26338572afca1063b99a1e52048925e9d52e43d38", "request_sha256": "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "compiled_tokens": 404, "compiled_prompt_sha256": "63461712757b659f4c1e81c96db955b9576be9a529dd979a61ae859b5e4fe20e", "compiled_ids_sha256": "82d8ecc5cf442b09faecb1a0ed1d65f9bc9153c91f2b0ea1a85d3471da004e30", "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": "2026-09-28T15:32:52Z"} {"mode": "kev-choice", "prompt_sha256": "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "n_options": 6, "kev_response": {"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": 172.22962899904815, "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": 0.16871440599788912, "probabilities_f64_minus_runtime_max_abs": 9.319700899901306e-09}}}, "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": 0.17225294899981236, "prompt_tokens": 309, "eval_tokens": 1, "server_latency_ms": 172.22962899904815, "prompt_eval_ms": null, "total_ms": 172.25294899981236, "done_reason": null, "id": "c-sagan-01-dessert-01", "uid": "c073", "task": "kc", "label": "sagan", "text_sha256": "af4421e88e6f7cc45a04171042c0ae408960b8d2dd13e1e0b67b4579a5dafe83", "request_sha256": "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "compiled_tokens": 309, "compiled_prompt_sha256": "929aa8fbc3eee78c7018d5c8c61cbd18f10472b9ec10356a479f5560f36dc969", "compiled_ids_sha256": "9bfb6bfa3765f21c286543e6595240c7937c405557de6095941783fa77c32cf4", "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": "2026-09-28T15:32:52Z"} {"mode": "kev-choice", "prompt_sha256": "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "n_options": 6, "kev_response": {"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": 169.49935000229743, "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": 0.16620901599890203, "probabilities_f64_minus_runtime_max_abs": 1.7307154043244566e-08}}}, "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": 0.16952120999849285, "prompt_tokens": 288, "eval_tokens": 1, "server_latency_ms": 169.49935000229743, "prompt_eval_ms": null, "total_ms": 169.52120999849285, "done_reason": null, "id": "c-sagan-01-dessert-01", "uid": "c074", "task": "kc", "label": "sagan", "text_sha256": "201a0135be6be1db753dcac7891bcbd6ebee66569e3359c08b3ee0a4105fd3fd", "request_sha256": "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "compiled_tokens": 288, "compiled_prompt_sha256": "c9162970d7a87671df926d763cd3c86cd736e6a98d9885f5541c96e5393b7873", "compiled_ids_sha256": "2e18cf1277f24da4cbb35e3c4d9ff0025b081d2a66b6e030656f384eefb90cf3", "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": "2026-09-28T15:32:52Z"} {"mode": "kev-choice", "prompt_sha256": "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "n_options": 6, "kev_response": {"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": 179.7221009983332, "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": 0.176047478998953, "probabilities_f64_minus_runtime_max_abs": 4.27186938045665e-08}}}, "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": 0.17974781099837855, "prompt_tokens": 327, "eval_tokens": 1, "server_latency_ms": 179.7221009983332, "prompt_eval_ms": null, "total_ms": 179.74781099837855, "done_reason": null, "id": "c-sagan-01-dessert-03", "uid": "c075", "task": "kc", "label": "sagan", "text_sha256": "b17461d357df5f854916b406f705dfee2535c04fe6e0ec22055fb868d847f9a2", "request_sha256": "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "compiled_tokens": 327, "compiled_prompt_sha256": "efc2df06e0745877c07b202c66d4cf238051ae0a1ae157bb1d319d831d699748", "compiled_ids_sha256": "6a8b4cc8d5bca66563e796c67461a643efc58f4871ab7302ea9ea36cd395825a", "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": "2026-09-28T15:32:53Z"} {"mode": "kev-choice", "prompt_sha256": "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "n_options": 6, "kev_response": {"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": 208.2895079984155, "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": 0.20347575599953416, "probabilities_f64_minus_runtime_max_abs": 1.7772009530681032e-08}}}, "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": 0.20831633800116833, "prompt_tokens": 409, "eval_tokens": 1, "server_latency_ms": 208.2895079984155, "prompt_eval_ms": null, "total_ms": 208.31633800116833, "done_reason": null, "id": "c-sagan-01-dessert-03", "uid": "c076", "task": "kc", "label": "sagan", "text_sha256": "0a559d120c34e49e361afce6d91c183b88fe89ed3166d332e2a0801f986305ea", "request_sha256": "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "compiled_tokens": 409, "compiled_prompt_sha256": "d1ab3fa14e7acc3c64152e9b9fefb12997ba7cb18617a7d4ba7df9edddd170aa", "compiled_ids_sha256": "b4fb76f1da222baeff1f762980a440de6c084e5a94b0a0af875b6114f77bd052", "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": "2026-09-28T15:32:53Z"} {"mode": "kev-choice", "prompt_sha256": "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "n_options": 6, "kev_response": {"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": 180.43357499846024, "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": 0.17677310300132376, "probabilities_f64_minus_runtime_max_abs": 2.6277877096525515e-08}}}, "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": 0.1804570150015934, "prompt_tokens": 331, "eval_tokens": 1, "server_latency_ms": 180.43357499846024, "prompt_eval_ms": null, "total_ms": 180.4570150015934, "done_reason": null, "id": "c-sagan-01-dessert-03", "uid": "c077", "task": "kc", "label": "sagan", "text_sha256": "0a4333ec1c14f2d7026dd16373b6b7f5850ff30627e3985ba2719899cee3426f", "request_sha256": "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "compiled_tokens": 331, "compiled_prompt_sha256": "5323ed669bcbd837510d737e7f7ff27ae1e5464e1a1530c9bd5f04eaf16e3def", "compiled_ids_sha256": "9fa1f31f937a41e05c35b71c72dada3a715a34d8f8cc835e0cc5cabd5ccf3784", "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": "2026-09-28T15:32:53Z"} {"mode": "kev-choice", "prompt_sha256": "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "n_options": 6, "kev_response": {"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": 188.0933249994996, "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": 0.18397081800139858, "probabilities_f64_minus_runtime_max_abs": 2.74611841044603e-08}}}, "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": 0.18811846500102547, "prompt_tokens": 360, "eval_tokens": 1, "server_latency_ms": 188.0933249994996, "prompt_eval_ms": null, "total_ms": 188.11846500102547, "done_reason": null, "id": "c-sagan-01-dessert-03", "uid": "c078", "task": "kc", "label": "sagan", "text_sha256": "4adaea24337f52c5fd58eb48284a3be16efee7fd902865cb45b5695012a24e94", "request_sha256": "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "compiled_tokens": 360, "compiled_prompt_sha256": "9185a3cc0cf1abb59a3142933ef89d11e19cf783575408d4218a74be6254ce2d", "compiled_ids_sha256": "36b52d662b922ffae58003f7a5d011a210ce57c6309ab056b16d4638019782b0", "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": "2026-09-28T15:32:53Z"} {"mode": "kev-choice", "prompt_sha256": "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "n_options": 6, "kev_response": {"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": 206.32175400169217, "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": 0.20196193699666765, "probabilities_f64_minus_runtime_max_abs": 5.538823588668862e-08}}}, "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": 0.20634832400173764, "prompt_tokens": 397, "eval_tokens": 1, "server_latency_ms": 206.32175400169217, "prompt_eval_ms": null, "total_ms": 206.34832400173764, "done_reason": null, "id": "c-sagan-01-dessert-03", "uid": "c079", "task": "kc", "label": "sagan", "text_sha256": "8c630a6e410940f6d22264df1640a7a2f0f4fe19eccbd45cbb1a2fee09efa8c2", "request_sha256": "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "compiled_tokens": 397, "compiled_prompt_sha256": "4c9fb986381beb2b96309a2afbd1b2af9fe10613bf4378bdb77d7479f8c41680", "compiled_ids_sha256": "86437e7858dab9217d982273312aa4f196d2a66f6de0c27248ed494a7c977df0", "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": "2026-09-28T15:32:53Z"} {"mode": "kev-choice", "prompt_sha256": "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "n_options": 6, "kev_response": {"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": 180.1011669995205, "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": 0.17634510700008832, "probabilities_f64_minus_runtime_max_abs": 4.511059525924566e-09}}}, "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": 0.18012565699973493, "prompt_tokens": 333, "eval_tokens": 1, "server_latency_ms": 180.1011669995205, "prompt_eval_ms": null, "total_ms": 180.12565699973493, "done_reason": null, "id": "c-sagan-01-main-01", "uid": "c080", "task": "kc", "label": "sagan", "text_sha256": "70f507eb1365d078f2ec073a7217f16108c3aa295d0f6f410b17376071934378", "request_sha256": "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "compiled_tokens": 333, "compiled_prompt_sha256": "9039ec9ebd536e8ff815ec4d3af1f3487d4b1b3ac68f56b8c74d168136f644a8", "compiled_ids_sha256": "d5b1938a23c4d5daa57a1a68304eaa2e03f3dfd41e38df0bac9ead15b4c8f182", "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": "2026-09-28T15:32:54Z"} {"mode": "kev-choice", "prompt_sha256": "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "n_options": 6, "kev_response": {"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": 190.87336400116328, "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": 0.18640649799999665, "probabilities_f64_minus_runtime_max_abs": 2.6528217067500748e-08}}}, "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": 0.19090316399888252, "prompt_tokens": 366, "eval_tokens": 1, "server_latency_ms": 190.87336400116328, "prompt_eval_ms": null, "total_ms": 190.90316399888252, "done_reason": null, "id": "c-sagan-01-main-02", "uid": "c081", "task": "kc", "label": "sagan", "text_sha256": "26c9cc6cd1f5dfbf4575a10bfddada00c88453f4871b30ab1fc5c9322fa24311", "request_sha256": "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "compiled_tokens": 366, "compiled_prompt_sha256": "e899284aeb4ab20e1a02f56556620f4fbd0c64a5bfca3086c98d666d6ed9d79b", "compiled_ids_sha256": "baf3a4db52592f705e8840880143e8aa8842446449b21f0d838487828fea0bfd", "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": "2026-09-28T15:32:54Z"} {"mode": "kev-choice", "prompt_sha256": "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "n_options": 6, "kev_response": {"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": 172.88887399990926, "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": 0.16926806199990097, "probabilities_f64_minus_runtime_max_abs": 7.157135328839104e-09}}}, "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": 0.17291264399682404, "prompt_tokens": 316, "eval_tokens": 1, "server_latency_ms": 172.88887399990926, "prompt_eval_ms": null, "total_ms": 172.91264399682404, "done_reason": null, "id": "c-sagan-01-main-02", "uid": "c082", "task": "kc", "label": "sagan", "text_sha256": "090da8cbfb7483c7bf328f01ab049ab77c06781715061f8708fd5d7b7edf51bc", "request_sha256": "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "compiled_tokens": 316, "compiled_prompt_sha256": "6b3262c0b05359df640a75374ab1cb9ee093c35d018b01ad37614f989d8ad180", "compiled_ids_sha256": "9cf18e6fda3d59410ee91dbb91dadbf48ceba6733de0f541fd3d4d6902aa5f4c", "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": "2026-09-28T15:32:54Z"} {"mode": "kev-choice", "prompt_sha256": "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "n_options": 6, "kev_response": {"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": 206.12705500025186, "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": 0.2016782200007583, "probabilities_f64_minus_runtime_max_abs": 3.27161010504895e-08}}}, "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": 0.20615325499966275, "prompt_tokens": 399, "eval_tokens": 1, "server_latency_ms": 206.12705500025186, "prompt_eval_ms": null, "total_ms": 206.15325499966275, "done_reason": null, "id": "c-sagan-01-main-03", "uid": "c083", "task": "kc", "label": "sagan", "text_sha256": "4ff2c7c47e30b6204ebc63d53fa463e60b4fc427332819e30a966cd466e5cf29", "request_sha256": "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "compiled_tokens": 399, "compiled_prompt_sha256": "24b38735ab28fcd79f74277c7c0d9977f3ca86b28f20b968104f2e28ecd092e0", "compiled_ids_sha256": "c413e920def965f2042bf7df2f36f935761ed11a9ea1a07f8d6b709f90857e2d", "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": "2026-09-28T15:32:54Z"} {"mode": "kev-choice", "prompt_sha256": "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "n_options": 6, "kev_response": {"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": 207.19962700241012, "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": 0.20258789299987257, "probabilities_f64_minus_runtime_max_abs": 4.1708198783574346e-08}}}, "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": 0.20722770599968499, "prompt_tokens": 400, "eval_tokens": 1, "server_latency_ms": 207.19962700241012, "prompt_eval_ms": null, "total_ms": 207.22770599968499, "done_reason": null, "id": "c-sagan-01-main-03", "uid": "c084", "task": "kc", "label": "sagan", "text_sha256": "75fd209542847cbf68001d0bfbdc0906a9af020a3cae869a8c521fa0d7733d54", "request_sha256": "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "compiled_tokens": 400, "compiled_prompt_sha256": "18fc51ab32b6e239ab8298165343e7ff14bc85e0d166e6dc0b5e9964cae0940b", "compiled_ids_sha256": "73adbcdaa744c056b6a9cf97351cb9d64087fb324180be15d21ff25d3141cbfd", "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": "2026-09-28T15:32:54Z"} {"mode": "kev-choice", "prompt_sha256": "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "n_options": 6, "kev_response": {"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": 169.25856200032285, "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": 0.1661109070009843, "probabilities_f64_minus_runtime_max_abs": 4.005959208797094e-08}}}, "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": 0.16928124200057937, "prompt_tokens": 284, "eval_tokens": 1, "server_latency_ms": 169.25856200032285, "prompt_eval_ms": null, "total_ms": 169.28124200057937, "done_reason": null, "id": "c-sagan-01-starter-01", "uid": "c085", "task": "kc", "label": "sagan", "text_sha256": "75430e1e6218b85b7b5a24b15accfed19fcb961b965741f729b8451ae554e6c3", "request_sha256": "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "compiled_tokens": 284, "compiled_prompt_sha256": "65b093d4dc7322a71d795f378d96f989360cca3b9cc8cc2dafc040037e482792", "compiled_ids_sha256": "c711190daf3352b61206e8801960f56744ba55154d91e3a74652ef38d6ba1604", "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": "2026-09-28T15:32:55Z"} {"mode": "kev-choice", "prompt_sha256": "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "n_options": 6, "kev_response": {"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": 206.92450900241965, "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": 0.20248213400191162, "probabilities_f64_minus_runtime_max_abs": 9.032790515561828e-09}}}, "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": 0.20694984899819246, "prompt_tokens": 404, "eval_tokens": 1, "server_latency_ms": 206.92450900241965, "prompt_eval_ms": null, "total_ms": 206.94984899819246, "done_reason": null, "id": "c-sagan-01-starter-02", "uid": "c086", "task": "kc", "label": "sagan", "text_sha256": "75df625f87a54ffae2b7b0cdadfa23d16456e82b7c6e3d061a0358a590f5043f", "request_sha256": "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "compiled_tokens": 404, "compiled_prompt_sha256": "e2b81005504d2eebe6b4f6eae69c57ba10486928d37eddc53c160a4b2854c594", "compiled_ids_sha256": "6f3ebc0833ff0d4eca3500ecdd2bf006ccd0dd84cbb8ec6be2d67191859098b9", "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": "2026-09-28T15:32:55Z"} {"mode": "kev-choice", "prompt_sha256": "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "n_options": 6, "kev_response": {"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": 170.01006600185065, "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": 0.16668490300071426, "probabilities_f64_minus_runtime_max_abs": 1.3553915145791962e-07}}}, "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": 0.17003459600164206, "prompt_tokens": 293, "eval_tokens": 1, "server_latency_ms": 170.01006600185065, "prompt_eval_ms": null, "total_ms": 170.03459600164206, "done_reason": null, "id": "c-sagan-01-starter-02", "uid": "c087", "task": "kc", "label": "sagan", "text_sha256": "f37ad2a441d669b186a45bde189d7d209f5077324c7b15113285fce54185d55b", "request_sha256": "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "compiled_tokens": 293, "compiled_prompt_sha256": "260bb13ff919a69d62276bf87d96d93384b67dfa42212c818966dc0bd3ee540c", "compiled_ids_sha256": "56677b4e63c956b87e199aa8482bbf7c197f78ec3d9d566703cd84fe0d74f83f", "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": "2026-09-28T15:32:55Z"} {"mode": "kev-choice", "prompt_sha256": "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "n_options": 6, "kev_response": {"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": 170.85641999801737, "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": 0.16748768600155017, "probabilities_f64_minus_runtime_max_abs": 3.8691211456942654e-08}}}, "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": 0.17087933999937377, "prompt_tokens": 302, "eval_tokens": 1, "server_latency_ms": 170.85641999801737, "prompt_eval_ms": null, "total_ms": 170.87933999937377, "done_reason": null, "id": "c-sagan-01-starter-03", "uid": "c088", "task": "kc", "label": "sagan", "text_sha256": "36439125e5a22ecbd8e1ade7eb23aeea2e65ec5efd46bd993882d91dd00c6cf1", "request_sha256": "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "compiled_tokens": 302, "compiled_prompt_sha256": "3824b179b4488ce489b4ef53aaa0bc5aac8823a7a8d88bcb5a7f0cf96ad47e1b", "compiled_ids_sha256": "4010bc1586d12cb94c392181608c684fece802cc25b6f26e7c35ffddfc2a80eb", "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": "2026-09-28T15:32:55Z"} {"mode": "kev-choice", "prompt_sha256": "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "n_options": 6, "kev_response": {"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": 182.59632800254622, "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": 0.1786402489997272, "probabilities_f64_minus_runtime_max_abs": 3.4965003004039374e-08}}}, "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": 0.1826237780005613, "prompt_tokens": 355, "eval_tokens": 1, "server_latency_ms": 182.59632800254622, "prompt_eval_ms": null, "total_ms": 182.6237780005613, "done_reason": null, "id": "c-sagan-01-starter-03", "uid": "c089", "task": "kc", "label": "sagan", "text_sha256": "b82156a1523f861c6acec464a68044a0347260585b05e3d579f9d1bb6d3e1883", "request_sha256": "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "compiled_tokens": 355, "compiled_prompt_sha256": "ce02283a565951bd24fed4ecdd1be1ff580fd4e71e1ed42c0be57924ef7735f0", "compiled_ids_sha256": "fcc59376315f837186e1531a6bc1a77307e5878970fb6cc3e97bcfa34479b565", "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": "2026-09-28T15:32:55Z"} {"mode": "kev-choice", "prompt_sha256": "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "n_options": 6, "kev_response": {"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": 207.61053300157073, "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": 0.2029085700014548, "probabilities_f64_minus_runtime_max_abs": 1.2188878462104924e-08}}}, "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": 0.20763995299785165, "prompt_tokens": 409, "eval_tokens": 1, "server_latency_ms": 207.61053300157073, "prompt_eval_ms": null, "total_ms": 207.63995299785165, "done_reason": null, "id": "c-socrates-01-dessert-01", "uid": "c090", "task": "kc", "label": "socrates", "text_sha256": "b63520cf2b49ef5a68364a7ce70d03329569ecdd3852bcdf74b6daa63c3968bf", "request_sha256": "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "compiled_tokens": 409, "compiled_prompt_sha256": "e5874d1b8d5df9f9ce009834e203a80ed9bfc65e1e117e65b805719bfafff002", "compiled_ids_sha256": "2fbf7e93347627cacd15188db8ccc02538cff99269ff3adb9d0ea548870e9cb9", "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": "2026-09-28T15:32:56Z"} {"mode": "kev-choice", "prompt_sha256": "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "n_options": 6, "kev_response": {"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": 169.90640700169024, "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": 0.16668553199997405, "probabilities_f64_minus_runtime_max_abs": 2.201190230088912e-08}}}, "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": 0.1699287869996624, "prompt_tokens": 291, "eval_tokens": 1, "server_latency_ms": 169.90640700169024, "prompt_eval_ms": null, "total_ms": 169.9287869996624, "done_reason": null, "id": "c-socrates-01-dessert-01", "uid": "c091", "task": "kc", "label": "socrates", "text_sha256": "f537b42e13f651d15d8bcfb648d10c7339118d0af15ef233f9a4976b21b55185", "request_sha256": "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "compiled_tokens": 291, "compiled_prompt_sha256": "38c5218cbd49d0ad53554b844ea9581c36f8d9f6dab540ad7633c7aefaaec7f9", "compiled_ids_sha256": "2b8caa07c5125e82419f0a570ae0c7c4b742c0c93eaa8a8cd7f945e8e8e0643e", "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": "2026-09-28T15:32:56Z"} {"mode": "kev-choice", "prompt_sha256": "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "n_options": 6, "kev_response": {"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": 187.62040000001434, "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": 0.1836391700016975, "probabilities_f64_minus_runtime_max_abs": 7.518704547848643e-09}}}, "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": 0.18765150899707805, "prompt_tokens": 361, "eval_tokens": 1, "server_latency_ms": 187.62040000001434, "prompt_eval_ms": null, "total_ms": 187.65150899707805, "done_reason": null, "id": "c-socrates-01-dessert-01", "uid": "c092", "task": "kc", "label": "socrates", "text_sha256": "58f8bc29dfc873fb65ccddb7a73ce568f962dc7f1f25e88e12dd0b0798caf239", "request_sha256": "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "compiled_tokens": 361, "compiled_prompt_sha256": "9a53e19f51ada7e3ca7748e01853f6e0dfe0321fc114aca34ef2b3bce186324a", "compiled_ids_sha256": "08a973dfb69a66733142afb8c93b88554f14300d2455d7ab0777f84c253ecdf1", "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": "2026-09-28T15:32:56Z"} {"mode": "kev-choice", "prompt_sha256": "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "n_options": 6, "kev_response": {"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": 180.9915909980191, "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": 0.17718248099845368, "probabilities_f64_minus_runtime_max_abs": 7.09965369827259e-09}}}, "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": 0.18101557000045432, "prompt_tokens": 343, "eval_tokens": 1, "server_latency_ms": 180.9915909980191, "prompt_eval_ms": null, "total_ms": 181.01557000045432, "done_reason": null, "id": "c-socrates-01-dessert-02", "uid": "c093", "task": "kc", "label": "socrates", "text_sha256": "6b8746afe3db64759c4fe8e62f4167572b8707c6e0af2a303917c790e0107cf7", "request_sha256": "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "compiled_tokens": 343, "compiled_prompt_sha256": "98bf941e9d49195d9f221459067344330d3de649eeabf47b96667525ca3aa283", "compiled_ids_sha256": "0f0f1ae75907c024b2d459061eabb496925526ca0d95c76d438b65fd51aa1e6e", "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": "2026-09-28T15:32:56Z"} {"mode": "kev-choice", "prompt_sha256": "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "n_options": 6, "kev_response": {"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": 189.96162100302172, "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": 0.1858890830008022, "probabilities_f64_minus_runtime_max_abs": 1.815291650331119e-08}}}, "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": 0.1899861010024324, "prompt_tokens": 374, "eval_tokens": 1, "server_latency_ms": 189.96162100302172, "prompt_eval_ms": null, "total_ms": 189.9861010024324, "done_reason": null, "id": "c-socrates-01-dessert-02", "uid": "c094", "task": "kc", "label": "socrates", "text_sha256": "91e20cb4e96bf48056541da6446b16d67d78651938fceace8dbfebd07fb6d458", "request_sha256": "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "compiled_tokens": 374, "compiled_prompt_sha256": "e0d5ffa591269263ed024ca529109dfb26c248d9d18c4a8d6e57f2ba9168b233", "compiled_ids_sha256": "42ef38f2cbdef5189c3d759a7044b4c8a748d18138cc275d8d40d3f3af6771d4", "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": "2026-09-28T15:32:56Z"} {"mode": "kev-choice", "prompt_sha256": "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "n_options": 6, "kev_response": {"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": 170.43495300094946, "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": 0.16707821899763076, "probabilities_f64_minus_runtime_max_abs": 3.408563509221807e-08}}}, "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": 0.1704581930025597, "prompt_tokens": 302, "eval_tokens": 1, "server_latency_ms": 170.43495300094946, "prompt_eval_ms": null, "total_ms": 170.4581930025597, "done_reason": null, "id": "c-socrates-01-dessert-02", "uid": "c095", "task": "kc", "label": "socrates", "text_sha256": "9ffa8e8f85216a9cb308e9ac582dff3a3bef39e1d20109b23060822860780631", "request_sha256": "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "compiled_tokens": 302, "compiled_prompt_sha256": "3b30247295b5c3dec8716911428c2b205adc94f0d63fb7499beab302bd9b5b24", "compiled_ids_sha256": "0da5ad6c304119d9f5f455fa27621e3de8086264734e81f84f183030922025e2", "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": "2026-09-28T15:32:56Z"} {"mode": "kev-choice", "prompt_sha256": "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "n_options": 6, "kev_response": {"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": 179.8092600001837, "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": 0.17602518900093855, "probabilities_f64_minus_runtime_max_abs": 3.3477038585694174e-08}}}, "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": 0.17983312000069418, "prompt_tokens": 323, "eval_tokens": 1, "server_latency_ms": 179.8092600001837, "prompt_eval_ms": null, "total_ms": 179.83312000069418, "done_reason": null, "id": "c-socrates-01-main-01", "uid": "c096", "task": "kc", "label": "socrates", "text_sha256": "cdd1c49d12c260c7ebc9e05cb7f856333784a272c361416de516f780ca99b563", "request_sha256": "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "compiled_tokens": 323, "compiled_prompt_sha256": "1370e6b93e0997fcf22f6ff3e6473a4bccf01f168816a3e2da10413c14e6b44c", "compiled_ids_sha256": "a31f962774edf9fb5117efe523b6eb2f11ca98ce116e6d7c52e3734ea7a26aa9", "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": "2026-09-28T15:32:57Z"} {"mode": "kev-choice", "prompt_sha256": "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "n_options": 6, "kev_response": {"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": 205.80063799934578, "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": 0.20154147100038244, "probabilities_f64_minus_runtime_max_abs": 2.1215319834233526e-08}}}, "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": 0.20582560800175997, "prompt_tokens": 395, "eval_tokens": 1, "server_latency_ms": 205.80063799934578, "prompt_eval_ms": null, "total_ms": 205.82560800175997, "done_reason": null, "id": "c-socrates-01-main-02", "uid": "c097", "task": "kc", "label": "socrates", "text_sha256": "cde59da3738cd9a8fe6df261efdaf3a4456b8bc09d08411d3000ce3a29c83b5b", "request_sha256": "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "compiled_tokens": 395, "compiled_prompt_sha256": "7a7cb526fa35845664c998ea567eaa951e2590b39bd2a7c881e769dec476a428", "compiled_ids_sha256": "fba0ef685096385bfc59e52dd6d585816ee5d2066e5bbea9c00a20c6d1da8968", "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": "2026-09-28T15:32:57Z"} {"mode": "kev-choice", "prompt_sha256": "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "n_options": 6, "kev_response": {"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": 170.43062299853773, "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": 0.16709175900177797, "probabilities_f64_minus_runtime_max_abs": 3.521509456261285e-08}}}, "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": 0.17045324200080358, "prompt_tokens": 297, "eval_tokens": 1, "server_latency_ms": 170.43062299853773, "prompt_eval_ms": null, "total_ms": 170.45324200080358, "done_reason": null, "id": "c-socrates-01-main-02", "uid": "c098", "task": "kc", "label": "socrates", "text_sha256": "7fcce46b50361f48ee0b98a7e17832e9023126cbeccd0326a9cbb0727f43f7ad", "request_sha256": "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "compiled_tokens": 297, "compiled_prompt_sha256": "ae667f183d359d40c186509ad74566d963a401022b071a698cabca10ad469305", "compiled_ids_sha256": "455a8afb9852c21fd5125d01820fb2e27419fa8e5daf8123220881908a4db405", "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": "2026-09-28T15:32:57Z"} {"mode": "kev-choice", "prompt_sha256": "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "n_options": 6, "kev_response": {"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": 188.62590199933038, "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": 0.18461940199995297, "probabilities_f64_minus_runtime_max_abs": 1.267863836140748e-08}}}, "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": 0.18865036200077157, "prompt_tokens": 368, "eval_tokens": 1, "server_latency_ms": 188.62590199933038, "prompt_eval_ms": null, "total_ms": 188.65036200077157, "done_reason": null, "id": "c-socrates-01-main-02", "uid": "c099", "task": "kc", "label": "socrates", "text_sha256": "31b2c0bd67cd1c668ff7fbd043e45eec1df05d027787b72b24c896f2ef174f72", "request_sha256": "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "compiled_tokens": 368, "compiled_prompt_sha256": "241005cdbb0ac4de9754451ae5ce2a6d308deccb51feb8a49be378b2807af5d6", "compiled_ids_sha256": "7479577962a5718990d77f717d29e7cc80e2c6bc96a00d11d0d6eae9329107e2", "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": "2026-09-28T15:32:57Z"} {"mode": "kev-choice", "prompt_sha256": "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "n_options": 6, "kev_response": {"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": 170.51407200051472, "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": 0.16716537899992545, "probabilities_f64_minus_runtime_max_abs": 1.684721412376078e-08}}}, "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": 0.1705374720004329, "prompt_tokens": 298, "eval_tokens": 1, "server_latency_ms": 170.51407200051472, "prompt_eval_ms": null, "total_ms": 170.5374720004329, "done_reason": null, "id": "c-socrates-01-main-02", "uid": "c100", "task": "kc", "label": "socrates", "text_sha256": "16c378fcf97035a1f0d659bd939e505f9e11f421fe386b8d29c5bed28d5fb98c", "request_sha256": "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "compiled_tokens": 298, "compiled_prompt_sha256": "5a8c5648b9712c46350b863549c1ceed5801044f634662c32f1ac342b250aa4c", "compiled_ids_sha256": "b7bb323dfd3806fa1b596523b49ddfd4bc24913a34b61cf4015eef0afa9175a5", "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": "2026-09-28T15:32:57Z"} {"mode": "kev-choice", "prompt_sha256": "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "n_options": 6, "kev_response": {"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": 180.81589199937298, "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": 0.17709641200053738, "probabilities_f64_minus_runtime_max_abs": 3.053807112740614e-08}}}, "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": 0.18084001199895283, "prompt_tokens": 340, "eval_tokens": 1, "server_latency_ms": 180.81589199937298, "prompt_eval_ms": null, "total_ms": 180.84001199895283, "done_reason": null, "id": "c-socrates-01-main-03", "uid": "c101", "task": "kc", "label": "socrates", "text_sha256": "3f15b462d9b469045459958c57da06ec95d9f700f24d9477fff5a9b50ddd72aa", "request_sha256": "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "compiled_tokens": 340, "compiled_prompt_sha256": "aa8a76f665cd82962eddc0c1cfe089528415b48ad3760ba5f5715122b5f5c9b1", "compiled_ids_sha256": "86a40936a666d5fd48b556179b1723278e9807945ccc3cb11a9396580f0dd5a4", "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": "2026-09-28T15:32:58Z"} {"mode": "kev-choice", "prompt_sha256": "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "n_options": 6, "kev_response": {"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": 190.77751499935403, "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": 0.18644947800203227, "probabilities_f64_minus_runtime_max_abs": 1.625101397895179e-08}}}, "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": 0.19080242500058375, "prompt_tokens": 380, "eval_tokens": 1, "server_latency_ms": 190.77751499935403, "prompt_eval_ms": null, "total_ms": 190.80242500058375, "done_reason": null, "id": "c-socrates-01-main-03", "uid": "c102", "task": "kc", "label": "socrates", "text_sha256": "90e4b3b6a658a07636bbe827abfb381f93c6fb195ffd5ef085bf583f55af1d74", "request_sha256": "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "compiled_tokens": 380, "compiled_prompt_sha256": "302cd7d856db5d188e298657e794f2abd5a5c6239133269be866a306600f9f17", "compiled_ids_sha256": "064b264cbea6255d0964374c8390710e5279938b074c351063f0f26916c2824f", "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": "2026-09-28T15:32:58Z"} {"mode": "kev-choice", "prompt_sha256": "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "n_options": 6, "kev_response": {"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": 167.96846199940774, "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": 0.1648883060006483, "probabilities_f64_minus_runtime_max_abs": 1.2699127949211686e-07}}}, "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": 0.16799143199750688, "prompt_tokens": 267, "eval_tokens": 1, "server_latency_ms": 167.96846199940774, "prompt_eval_ms": null, "total_ms": 167.99143199750688, "done_reason": null, "id": "c-socrates-01-starter-01", "uid": "c103", "task": "kc", "label": "socrates", "text_sha256": "ce1f75337ffdd8f1d3134667d490376cc578923171abcb75d93c58afae704de3", "request_sha256": "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "compiled_tokens": 267, "compiled_prompt_sha256": "7ba348fbd49cae74d977fc56e0a7c6c7e8609ef11e84d162768dc1a823298c56", "compiled_ids_sha256": "279db28cb15e54380a75c8b3d192304b55a16372b39c8bf6c63f78d6061d4052", "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": "2026-09-28T15:32:58Z"} {"mode": "kev-choice", "prompt_sha256": "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "n_options": 6, "kev_response": {"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": 172.6786750004976, "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": 0.1691916229974595, "probabilities_f64_minus_runtime_max_abs": 3.8609576313852756e-08}}}, "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": 0.17270215499956976, "prompt_tokens": 315, "eval_tokens": 1, "server_latency_ms": 172.6786750004976, "prompt_eval_ms": null, "total_ms": 172.70215499956976, "done_reason": null, "id": "c-socrates-01-starter-02", "uid": "c104", "task": "kc", "label": "socrates", "text_sha256": "d2544f4722f7953ab187db9c08752f507f9e7bd525b074e990730d32763f93b0", "request_sha256": "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "compiled_tokens": 315, "compiled_prompt_sha256": "09a3150ed8b79aaef3b11c5a10df28e718cebf9189825e56bd6872be31be436d", "compiled_ids_sha256": "2fae6a509382acdcb1398799443b33dcf79f533516bd7b6f1ae007a1d10494dc", "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": "2026-09-28T15:32:58Z"} {"mode": "kev-choice", "prompt_sha256": "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "n_options": 6, "kev_response": {"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": 180.42443599915714, "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": 0.17679901399969822, "probabilities_f64_minus_runtime_max_abs": 1.6032371363827735e-07}}}, "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": 0.18044817500049248, "prompt_tokens": 335, "eval_tokens": 1, "server_latency_ms": 180.42443599915714, "prompt_eval_ms": null, "total_ms": 180.44817500049248, "done_reason": null, "id": "c-socrates-01-starter-03", "uid": "c105", "task": "kc", "label": "socrates", "text_sha256": "76bd6a400f7300d59d745ae87514d123122ad9184a8cbf280deb317001c2365e", "request_sha256": "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "compiled_tokens": 335, "compiled_prompt_sha256": "1b43313b82313432c0e61c8d24490e5ee29de32faa2f8898725bc668a978fa6f", "compiled_ids_sha256": "8635156608b5843ff3e2d0a3a4e35ad222d413b6cce00d96311035874e998319", "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": "2026-09-28T15:32:58Z"} {"mode": "kev-choice", "prompt_sha256": "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "n_options": 6, "kev_response": {"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": 182.45875000138767, "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": 0.1785470799986797, "probabilities_f64_minus_runtime_max_abs": 4.422106014700944e-08}}}, "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": 0.18248509899785859, "prompt_tokens": 355, "eval_tokens": 1, "server_latency_ms": 182.45875000138767, "prompt_eval_ms": null, "total_ms": 182.4850989978586, "done_reason": null, "id": "c-socrates-01-starter-03", "uid": "c106", "task": "kc", "label": "socrates", "text_sha256": "62ebf3000b7bc1337e703d293f168f2fef0e77979b30055c111974f6d491f990", "request_sha256": "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "compiled_tokens": 355, "compiled_prompt_sha256": "e89581a53f340a2aeeb9384cd75a6f0aefe0ea16e3efc03c033b3b5f3f2847e8", "compiled_ids_sha256": "67d6cefb26c2e1a4ee433d8b03c2659df0da773a4de7d34f23c8e3cc19bf6703", "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": "2026-09-28T15:32:58Z"} {"mode": "kev-choice", "prompt_sha256": "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "n_options": 6, "kev_response": {"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": 179.95928800155525, "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": 0.17626273800124181, "probabilities_f64_minus_runtime_max_abs": 1.0720173193590199e-08}}}, "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": 0.17998385800092365, "prompt_tokens": 328, "eval_tokens": 1, "server_latency_ms": 179.95928800155525, "prompt_eval_ms": null, "total_ms": 179.98385800092365, "done_reason": null, "id": "c-socrates-01-starter-03", "uid": "c107", "task": "kc", "label": "socrates", "text_sha256": "d9b9706fa1986a4b001053dac8268bae4e0be7b3860099d80287a3aab18de4e2", "request_sha256": "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "compiled_tokens": 328, "compiled_prompt_sha256": "8501d29a2806beb81cce7eb55e95f965f1f81dc43c69104f947a1d2360bdb71d", "compiled_ids_sha256": "da1348b4fbd3b223210453e72685938201ef08da043c35159863b5490ea36eae", "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": "2026-09-28T15:32:59Z"}