{"mode": "kev-choice", "prompt_sha256": "ada2e647fb31ad59f271b00730eac9606a376e7312144f898a5078a179b3253d", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0808568253852305, "einstein": 0.04327953986120903, "hypatia": 0.275680476934937, "ibn_sina": 0.2599914518068516, "sagan": 0.16080299831508055, "socrates": 0.17938870769669135}, "confidence": 0.08089151858070398, "choice": "hypatia"}}, "usage": {"input_tokens": 279, "output_tokens": 1}, "latency_ms": 115.31917500178679, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-1.984375, -2.609375, -0.7578125, -0.81640625, -1.296875, -1.1875], "runtime_probabilities": [0.0808568149805069, 0.043279536068439484, 0.27568045258522034, 0.25999143719673157, 0.16080300509929657, 0.17938868701457977], "decide_seconds": 0.10905446199831204, "probabilities_f64_minus_runtime_max_abs": 2.4349716687410705e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0808568253852305, "einstein": 0.04327953986120903, "hypatia": 0.275680476934937, "ibn_sina": 0.2599914518068516, "sagan": 0.16080299831508055, "socrates": 0.17938870769669135}, "confidence": 0.08089151858070398, "seconds": 0.11538195399771212, "prompt_tokens": 279, "eval_tokens": 1, "server_latency_ms": 115.31917500178679, "prompt_eval_ms": null, "total_ms": 115.38195399771212, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:56Z"}
{"mode": "kev-choice", "prompt_sha256": "ae89c5003c7b8c49cffb6cc5b1d1490d93f6203de46ad9f7f01410a034925e66", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.9293999680939337, "einstein": 0.00862661060521792, "hypatia": 0.008559477788883278, "ibn_sina": 0.009699159015532944, "sagan": 0.021857392248216104, "socrates": 0.021857392248216104}, "confidence": 0.7980245944290134, "choice": "darwin"}}, "usage": {"input_tokens": 379, "output_tokens": 1}, "latency_ms": 95.45500799868023, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [2.703125, -1.9765625, -1.984375, -1.859375, -1.046875, -1.046875], "runtime_probabilities": [0.9293999671936035, 0.008626610971987247, 0.008559477515518665, 0.009699159301817417, 0.02185739204287529, 0.02185739204287529], "decide_seconds": 0.09074616400175728, "probabilities_f64_minus_runtime_max_abs": 9.00330143771555e-10}}}, "choice": "darwin", "probs": {"darwin": 0.9293999680939337, "einstein": 0.00862661060521792, "hypatia": 0.008559477788883278, "ibn_sina": 0.009699159015532944, "sagan": 0.021857392248216104, "socrates": 0.021857392248216104}, "confidence": 0.7980245944290134, "seconds": 0.09548426699984702, "prompt_tokens": 379, "eval_tokens": 1, "server_latency_ms": 95.45500799868023, "prompt_eval_ms": null, "total_ms": 95.48426699984702, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:56Z"}
{"mode": "kev-choice", "prompt_sha256": "6be4957f07feb5d82a63ae1dca58d96723d78d0c8f0e1a49f0cfafc844c756bc", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.036058562787407086, "einstein": 0.06529362210482342, "hypatia": 0.3434531151660357, "ibn_sina": 0.21916653197826416, "sagan": 0.1426143823424439, "socrates": 0.1934137856210258}, "confidence": 0.11079855299013008, "choice": "hypatia"}}, "usage": {"input_tokens": 296, "output_tokens": 1}, "latency_ms": 91.1832699966908, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.21875, -2.625, -0.96484375, -1.4140625, -1.84375, -1.5390625], "runtime_probabilities": [0.03605856001377106, 0.06529361754655838, 0.3434531092643738, 0.21916653215885162, 0.14261437952518463, 0.19341379404067993], "decide_seconds": 0.08748747899880982, "probabilities_f64_minus_runtime_max_abs": 8.419654140556787e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.036058562787407086, "einstein": 0.06529362210482342, "hypatia": 0.3434531151660357, "ibn_sina": 0.21916653197826416, "sagan": 0.1426143823424439, "socrates": 0.1934137856210258}, "confidence": 0.11079855299013008, "seconds": 0.09120765999978175, "prompt_tokens": 296, "eval_tokens": 1, "server_latency_ms": 91.1832699966908, "prompt_eval_ms": null, "total_ms": 91.20765999978175, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:56Z"}
{"mode": "kev-choice", "prompt_sha256": "1e254c48e8d558d49c1284459198465a15cef2148c47d5875cf4a5f35404bff2", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.11333382569507963, "einstein": 0.047244567477142924, "hypatia": 0.33571922585957326, "ibn_sina": 0.1728134472888929, "sagan": 0.10899207501494554, "socrates": 0.22189685866436576}, "confidence": 0.08667909312798061, "choice": "hypatia"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 97.81520999968052, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-1.953125, -2.828125, -0.8671875, -1.53125, -1.9921875, -1.28125], "runtime_probabilities": [0.1133338212966919, 0.047244563698768616, 0.3357192277908325, 0.17281344532966614, 0.1089920699596405, 0.22189685702323914], "decide_seconds": 0.09301441600109683, "probabilities_f64_minus_runtime_max_abs": 5.055305035694957e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.11333382569507963, "einstein": 0.047244567477142924, "hypatia": 0.33571922585957326, "ibn_sina": 0.1728134472888929, "sagan": 0.10899207501494554, "socrates": 0.22189685866436576}, "confidence": 0.08667909312798061, "seconds": 0.09784387000036077, "prompt_tokens": 399, "eval_tokens": 1, "server_latency_ms": 97.81520999968052, "prompt_eval_ms": null, "total_ms": 97.84387000036077, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:56Z"}
{"mode": "kev-choice", "prompt_sha256": "7850492fc6179b1b4b4bdd20394ab04399c688ce7f0b9d2de32e671e0627905c", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.007641598985714103, "einstein": 0.02883909329005066, "hypatia": 0.5725002216580144, "ibn_sina": 0.12527131092581492, "sagan": 0.1476056768697445, "socrates": 0.11814209827066142}, "confidence": 0.300253578612777, "choice": "hypatia"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 98.16513700206997, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.671875, -2.34375, 0.64453125, -0.875, -0.7109375, -0.93359375], "runtime_probabilities": [0.0076415990479290485, 0.028839092701673508, 0.5725002288818359, 0.12527132034301758, 0.14760567247867584, 0.11814209073781967], "decide_seconds": 0.09325948499827064, "probabilities_f64_minus_runtime_max_abs": 9.417202656569756e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.007641598985714103, "einstein": 0.02883909329005066, "hypatia": 0.5725002216580144, "ibn_sina": 0.12527131092581492, "sagan": 0.1476056768697445, "socrates": 0.11814209827066142}, "confidence": 0.300253578612777, "seconds": 0.09819351600162918, "prompt_tokens": 409, "eval_tokens": 1, "server_latency_ms": 98.16513700206997, "prompt_eval_ms": null, "total_ms": 98.19351600162918, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:56Z"}
{"mode": "kev-choice", "prompt_sha256": "1ef9e9b47aa0b69dc706e2b2d978416f2bd011a8d4c23e06e7b821ffac0d6d06", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.9550966506660542, "einstein": 0.005859489831018035, "hypatia": 0.003724518345485541, "ibn_sina": 0.004563375268795779, "sagan": 0.01525784489677557, "socrates": 0.015498120991870841}, "confidence": 0.8616884039209008, "choice": "darwin"}}, "usage": {"input_tokens": 411, "output_tokens": 1}, "latency_ms": 98.68087300128536, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [3.25, -1.84375, -2.296875, -2.09375, -0.88671875, -0.87109375], "runtime_probabilities": [0.9550966620445251, 0.00585948908701539, 0.003724517999216914, 0.004563375376164913, 0.01525784470140934, 0.01549812126904726], "decide_seconds": 0.09364860200003022, "probabilities_f64_minus_runtime_max_abs": 1.1378470943235186e-08}}}, "choice": "darwin", "probs": {"darwin": 0.9550966506660542, "einstein": 0.005859489831018035, "hypatia": 0.003724518345485541, "ibn_sina": 0.004563375268795779, "sagan": 0.01525784489677557, "socrates": 0.015498120991870841}, "confidence": 0.8616884039209008, "seconds": 0.0987085130000196, "prompt_tokens": 411, "eval_tokens": 1, "server_latency_ms": 98.68087300128536, "prompt_eval_ms": null, "total_ms": 98.7085130000196, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:56Z"}
{"mode": "kev-choice", "prompt_sha256": "56bc6422a8c3ea07ea10491507d42c62f99c3b12a9c3b03c70a585fa6079a545", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.019514608440204754, "einstein": 0.07364736285926461, "hypatia": 0.3040697408772609, "ibn_sina": 0.2268498479574869, "sagan": 0.19786195407124577, "socrates": 0.1780564857945371}, "confidence": 0.109655291438847, "choice": "hypatia"}}, "usage": {"input_tokens": 314, "output_tokens": 1}, "latency_ms": 91.36334899812937, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.3125, -1.984375, -0.56640625, -0.859375, -0.99609375, -1.1015625], "runtime_probabilities": [0.01951460726559162, 0.07364736497402191, 0.30406975746154785, 0.22684983909130096, 0.1978619396686554, 0.1780564934015274], "decide_seconds": 0.08736630899875308, "probabilities_f64_minus_runtime_max_abs": 1.6584286943732707e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.019514608440204754, "einstein": 0.07364736285926461, "hypatia": 0.3040697408772609, "ibn_sina": 0.2268498479574869, "sagan": 0.19786195407124577, "socrates": 0.1780564857945371}, "confidence": 0.109655291438847, "seconds": 0.09139036899796338, "prompt_tokens": 314, "eval_tokens": 1, "server_latency_ms": 91.36334899812937, "prompt_eval_ms": null, "total_ms": 91.39036899796338, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:56Z"}
{"mode": "kev-choice", "prompt_sha256": "7d16a870429eb78dcdb3906b5987ef8ef9229dd0e68f08702c3d72a4601b76cf", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.00698379433184302, "einstein": 0.03491677489081448, "hypatia": 0.5724024738868385, "ibn_sina": 0.1252499223027681, "sagan": 0.12234851097761779, "socrates": 0.13809852361011804}, "confidence": 0.29577431129729115, "choice": "hypatia"}}, "usage": {"input_tokens": 295, "output_tokens": 1}, "latency_ms": 91.06740100105526, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.625, -2.015625, 0.78125, -0.73828125, -0.76171875, -0.640625], "runtime_probabilities": [0.006983794271945953, 0.034916773438453674, 0.5724024772644043, 0.1252499371767044, 0.12234851717948914, 0.13809853792190552], "decide_seconds": 0.08734404999995604, "probabilities_f64_minus_runtime_max_abs": 1.4873936304837798e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.00698379433184302, "einstein": 0.03491677489081448, "hypatia": 0.5724024738868385, "ibn_sina": 0.1252499223027681, "sagan": 0.12234851097761779, "socrates": 0.13809852361011804}, "confidence": 0.29577431129729115, "seconds": 0.09109190100207343, "prompt_tokens": 295, "eval_tokens": 1, "server_latency_ms": 91.06740100105526, "prompt_eval_ms": null, "total_ms": 91.09190100207343, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:56Z"}
{"mode": "kev-choice", "prompt_sha256": "99aa86d7673acbef6b356ef6e1e62bb359780b871a574e286cb0d67fa9e028ce", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.011149102726758437, "einstein": 0.025520550300199566, "hypatia": 0.5371938528998689, "ibn_sina": 0.13958905319311904, "sagan": 0.09593805969621343, "socrates": 0.19060938118384066}, "confidence": 0.2782290415908272, "choice": "hypatia"}}, "usage": {"input_tokens": 372, "output_tokens": 1}, "latency_ms": 94.34407600201666, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.0625, -2.234375, 0.8125, -0.53515625, -0.91015625, -0.2236328125], "runtime_probabilities": [0.01114910189062357, 0.025520550087094307, 0.5371938347816467, 0.13958905637264252, 0.0959380641579628, 0.1906093806028366], "decide_seconds": 0.08992896999916411, "probabilities_f64_minus_runtime_max_abs": 1.8118222144813956e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.011149102726758437, "einstein": 0.025520550300199566, "hypatia": 0.5371938528998689, "ibn_sina": 0.13958905319311904, "sagan": 0.09593805969621343, "socrates": 0.19060938118384066}, "confidence": 0.2782290415908272, "seconds": 0.09437170599994715, "prompt_tokens": 372, "eval_tokens": 1, "server_latency_ms": 94.34407600201666, "prompt_eval_ms": null, "total_ms": 94.37170599994715, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:57Z"}
{"mode": "kev-choice", "prompt_sha256": "e1f0f4be015ea8f0987af3df612d4f3a535cc0f500bba24dfada1308176486e1", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.9830752407723251, "einstein": 0.0029393417425911776, "hypatia": 0.0011510627651246926, "ibn_sina": 0.0014103115807090867, "sagan": 0.005078727099380104, "socrates": 0.006345316039869904}, "confidence": 0.9386640585862862, "choice": "darwin"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 98.06411700265016, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [4.25, -1.5625, -2.5, -2.296875, -1.015625, -0.79296875], "runtime_probabilities": [0.9830752611160278, 0.0029393420554697514, 0.0011510627809911966, 0.0014103116700425744, 0.005078727379441261, 0.006345316302031279], "decide_seconds": 0.09323632499945234, "probabilities_f64_minus_runtime_max_abs": 2.0343702722414037e-08}}}, "choice": "darwin", "probs": {"darwin": 0.9830752407723251, "einstein": 0.0029393417425911776, "hypatia": 0.0011510627651246926, "ibn_sina": 0.0014103115807090867, "sagan": 0.005078727099380104, "socrates": 0.006345316039869904}, "confidence": 0.9386640585862862, "seconds": 0.09809130700159585, "prompt_tokens": 407, "eval_tokens": 1, "server_latency_ms": 98.06411700265016, "prompt_eval_ms": null, "total_ms": 98.09130700159585, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:57Z"}
{"mode": "kev-choice", "prompt_sha256": "b336b2cd695a572a1bc843edc2813cd4c22436dec5ab3c120fd4d7ac582be285", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.29084221399845395, "einstein": 0.10421058552388993, "hypatia": 0.15952409277211713, "ibn_sina": 0.08983513998962873, "sagan": 0.18577545489609296, "socrates": 0.16981251281981724}, "confidence": 0.041207524951234475, "choice": "darwin"}}, "usage": {"input_tokens": 401, "output_tokens": 1}, "latency_ms": 98.50119399925461, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-0.1064453125, -1.1328125, -0.70703125, -1.28125, -0.5546875, -0.64453125], "runtime_probabilities": [0.290842205286026, 0.10421059280633926, 0.1595240980386734, 0.08983514457941055, 0.18577547371387482, 0.16981250047683716], "decide_seconds": 0.09364760199969169, "probabilities_f64_minus_runtime_max_abs": 1.8817781860525074e-08}}}, "choice": "darwin", "probs": {"darwin": 0.29084221399845395, "einstein": 0.10421058552388993, "hypatia": 0.15952409277211713, "ibn_sina": 0.08983513998962873, "sagan": 0.18577545489609296, "socrates": 0.16981251281981724}, "confidence": 0.041207524951234475, "seconds": 0.09853128399845446, "prompt_tokens": 401, "eval_tokens": 1, "server_latency_ms": 98.50119399925461, "prompt_eval_ms": null, "total_ms": 98.53128399845446, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:57Z"}
{"mode": "kev-choice", "prompt_sha256": "c8f7e1b2d641fcf994218fd2b5f6995d0ac1bf809d856c20e7ffbe37ee02cca5", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.12789960960641947, "einstein": 0.039621450711878115, "hypatia": 0.2799045861218877, "ibn_sina": 0.17618915394524806, "sagan": 0.1655143927422205, "socrates": 0.21087080687234622}, "confidence": 0.06283392308661984, "choice": "hypatia"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 99.33161799926893, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-1.265625, -2.4375, -0.482421875, -0.9453125, -1.0078125, -0.765625], "runtime_probabilities": [0.12789961695671082, 0.03962145000696182, 0.2799045741558075, 0.1761891394853592, 0.16551437973976135, 0.21087080240249634], "decide_seconds": 0.09397672900013276, "probabilities_f64_minus_runtime_max_abs": 1.4459888869611248e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.12789960960641947, "einstein": 0.039621450711878115, "hypatia": 0.2799045861218877, "ibn_sina": 0.17618915394524806, "sagan": 0.1655143927422205, "socrates": 0.21087080687234622}, "confidence": 0.06283392308661984, "seconds": 0.09936007799842628, "prompt_tokens": 402, "eval_tokens": 1, "server_latency_ms": 99.33161799926893, "prompt_eval_ms": null, "total_ms": 99.36007799842628, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:57Z"}
{"mode": "kev-choice", "prompt_sha256": "97f03142dc8d0f609da8620d37ed9899a73abb02cf6db98cd2403169e1a8f1d1", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.02549203329227279, "einstein": 0.06612130327027212, "hypatia": 0.39143179901056724, "ibn_sina": 0.22456096592145564, "sagan": 0.11335828692317813, "socrates": 0.17903561158225403}, "confidence": 0.14583069147383176, "choice": "hypatia"}}, "usage": {"input_tokens": 240, "output_tokens": 1}, "latency_ms": 87.12710100007826, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.59375, -1.640625, 0.1376953125, -0.41796875, -1.1015625, -0.64453125], "runtime_probabilities": [0.025492031127214432, 0.06612130254507065, 0.3914317786693573, 0.22456097602844238, 0.11335828900337219, 0.17903561890125275], "decide_seconds": 0.08395686600124463, "probabilities_f64_minus_runtime_max_abs": 2.0341209938656846e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.02549203329227279, "einstein": 0.06612130327027212, "hypatia": 0.39143179901056724, "ibn_sina": 0.22456096592145564, "sagan": 0.11335828692317813, "socrates": 0.17903561158225403}, "confidence": 0.14583069147383176, "seconds": 0.08715001100063091, "prompt_tokens": 240, "eval_tokens": 1, "server_latency_ms": 87.12710100007826, "prompt_eval_ms": null, "total_ms": 87.15001100063091, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:57Z"}
{"mode": "kev-choice", "prompt_sha256": "b845ff9ea26f72b9ea56ec38a158e913f20101382e818927c12cf1babe316511", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.9776228328986314, "einstein": 0.004154476569086517, "hypatia": 0.002294319864422888, "ibn_sina": 0.002330450145052464, "sagan": 0.005881681522364787, "socrates": 0.007716239000441907}, "confidence": 0.9214632590603397, "choice": "darwin"}}, "usage": {"input_tokens": 269, "output_tokens": 1}, "latency_ms": 90.76812399871415, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [4.65625, -0.8046875, -1.3984375, -1.3828125, -0.45703125, -0.185546875], "runtime_probabilities": [0.9776228070259094, 0.004154476802796125, 0.0022943196818232536, 0.0023304501082748175, 0.005881681572645903, 0.007716239430010319], "decide_seconds": 0.08720656099831103, "probabilities_f64_minus_runtime_max_abs": 2.5872721964859124e-08}}}, "choice": "darwin", "probs": {"darwin": 0.9776228328986314, "einstein": 0.004154476569086517, "hypatia": 0.002294319864422888, "ibn_sina": 0.002330450145052464, "sagan": 0.005881681522364787, "socrates": 0.007716239000441907}, "confidence": 0.9214632590603397, "seconds": 0.09079400399787119, "prompt_tokens": 269, "eval_tokens": 1, "server_latency_ms": 90.76812399871415, "prompt_eval_ms": null, "total_ms": 90.79400399787119, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:57Z"}
{"mode": "kev-choice", "prompt_sha256": "9a29b19b68b4226e1997d783910f8c1302786371d55a405853d4f64b71916564", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.39509259436078026, "einstein": 0.027902869375334417, "hypatia": 0.1959673920960642, "ibn_sina": 0.15932036169182237, "sagan": 0.11475415168334739, "socrates": 0.10696263079265138}, "confidence": 0.12581669927859418, "choice": "darwin"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 91.09422099936637, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-0.099609375, -2.75, -0.80078125, -1.0078125, -1.3359375, -1.40625], "runtime_probabilities": [0.39509257674217224, 0.027902869507670403, 0.19596737623214722, 0.15932035446166992, 0.11475414782762527, 0.10696262866258621], "decide_seconds": 0.08737697900141939, "probabilities_f64_minus_runtime_max_abs": 1.761860801696713e-08}}}, "choice": "darwin", "probs": {"darwin": 0.39509259436078026, "einstein": 0.027902869375334417, "hypatia": 0.1959673920960642, "ibn_sina": 0.15932036169182237, "sagan": 0.11475415168334739, "socrates": 0.10696263079265138}, "confidence": 0.12581669927859418, "seconds": 0.09111923100135755, "prompt_tokens": 294, "eval_tokens": 1, "server_latency_ms": 91.09422099936637, "prompt_eval_ms": null, "total_ms": 91.11923100135755, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:57Z"}
{"mode": "kev-choice", "prompt_sha256": "30a9396b56bc695ae23a32eff428110f7cb586e12fd3fb70164da328567b34bb", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.24568468483162423, "einstein": 0.04875761765306919, "hypatia": 0.2635811907440534, "ibn_sina": 0.14443057536313367, "sagan": 0.11425398910980936, "socrates": 0.1832919422983102}, "confidence": 0.06129880422051093, "choice": "hypatia"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 90.5490950026433, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-0.4140625, -2.03125, -0.34375, -0.9453125, -1.1796875, -0.70703125], "runtime_probabilities": [0.2456846684217453, 0.04875761643052101, 0.26358118653297424, 0.14443056285381317, 0.11425398290157318, 0.18329192698001862], "decide_seconds": 0.0870327320008073, "probabilities_f64_minus_runtime_max_abs": 1.6409878927925448e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.24568468483162423, "einstein": 0.04875761765306919, "hypatia": 0.2635811907440534, "ibn_sina": 0.14443057536313367, "sagan": 0.11425398910980936, "socrates": 0.1832919422983102}, "confidence": 0.06129880422051093, "seconds": 0.09057255500010797, "prompt_tokens": 273, "eval_tokens": 1, "server_latency_ms": 90.5490950026433, "prompt_eval_ms": null, "total_ms": 90.57255500010797, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:57Z"}
{"mode": "kev-choice", "prompt_sha256": "e31463776e2d63ca347866822a0a37ee28f8ec2812693c8947769b0b0af4d1d8", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0919475431390134, "einstein": 0.04516311578076722, "hypatia": 0.26814576480053925, "ibn_sina": 0.180024550122807, "sagan": 0.15278485130768496, "socrates": 0.26193417484918824}, "confidence": 0.07415416308121181, "choice": "hypatia"}}, "usage": {"input_tokens": 396, "output_tokens": 1}, "latency_ms": 97.70722000030219, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-1.7734375, -2.484375, -0.703125, -1.1015625, -1.265625, -0.7265625], "runtime_probabilities": [0.091947540640831, 0.0451631173491478, 0.26814576983451843, 0.18002454936504364, 0.15278485417366028, 0.26193419098854065], "decide_seconds": 0.09294408700225176, "probabilities_f64_minus_runtime_max_abs": 1.6139352410071695e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0919475431390134, "einstein": 0.04516311578076722, "hypatia": 0.26814576480053925, "ibn_sina": 0.180024550122807, "sagan": 0.15278485130768496, "socrates": 0.26193417484918824}, "confidence": 0.07415416308121181, "seconds": 0.09773375000077067, "prompt_tokens": 396, "eval_tokens": 1, "server_latency_ms": 97.70722000030219, "prompt_eval_ms": null, "total_ms": 97.73375000077067, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:57Z"}
{"mode": "kev-choice", "prompt_sha256": "d3c89f4f7c26fdb7c98e0288179dc1fda4d9fc4fe1736900021b79f3d0c9b169", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.10058053022401148, "einstein": 0.06544899927966873, "hypatia": 0.24507975848282995, "ibn_sina": 0.23385673923843506, "sagan": 0.1353458971075041, "socrates": 0.21968807566755064}, "confidence": 0.05259812119379326, "choice": "hypatia"}}, "usage": {"input_tokens": 383, "output_tokens": 1}, "latency_ms": 94.7208129982755, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-1.8515625, -2.28125, -0.9609375, -1.0078125, -1.5546875, -1.0703125], "runtime_probabilities": [0.10058053582906723, 0.06544899940490723, 0.24507977068424225, 0.23385675251483917, 0.13534589111804962, 0.21968808770179749], "decide_seconds": 0.09011641899996903, "probabilities_f64_minus_runtime_max_abs": 1.327640411497022e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.10058053022401148, "einstein": 0.06544899927966873, "hypatia": 0.24507975848282995, "ibn_sina": 0.23385673923843506, "sagan": 0.1353458971075041, "socrates": 0.21968807566755064}, "confidence": 0.05259812119379326, "seconds": 0.09474949300056323, "prompt_tokens": 383, "eval_tokens": 1, "server_latency_ms": 94.7208129982755, "prompt_eval_ms": null, "total_ms": 94.74949300056323, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:57Z"}
{"mode": "kev-choice", "prompt_sha256": "f3c3ff5751f493093f849d48e30c91c3ca66897d3ab98f8906f24bbf0c764bb3", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.021803017352822068, "einstein": 0.06928984584778376, "hypatia": 0.3057211374402231, "ibn_sina": 0.27836226333221264, "sagan": 0.15860603669992115, "socrates": 0.1662176993270372}, "confidence": 0.11987170873376285, "choice": "hypatia"}}, "usage": {"input_tokens": 373, "output_tokens": 1}, "latency_ms": 94.73620299831964, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.359375, -2.203125, -0.71875, -0.8125, -1.375, -1.328125], "runtime_probabilities": [0.02180301770567894, 0.06928984820842743, 0.30572113394737244, 0.2783622741699219, 0.15860603749752045, 0.16621769964694977], "decide_seconds": 0.09022812799958047, "probabilities_f64_minus_runtime_max_abs": 1.0837709230049342e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.021803017352822068, "einstein": 0.06928984584778376, "hypatia": 0.3057211374402231, "ibn_sina": 0.27836226333221264, "sagan": 0.15860603669992115, "socrates": 0.1662176993270372}, "confidence": 0.11987170873376285, "seconds": 0.09476323299895739, "prompt_tokens": 373, "eval_tokens": 1, "server_latency_ms": 94.73620299831964, "prompt_eval_ms": null, "total_ms": 94.76323299895739, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:58Z"}
{"mode": "kev-choice", "prompt_sha256": "b67cff6debb40976045fe29b9a21746ce0ee5f687719bd311e2c3fe3416d62fd", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0845542359698426, "einstein": 0.03416299956226512, "hypatia": 0.36300783725471714, "ibn_sina": 0.237138219141072, "sagan": 0.13617702234931925, "socrates": 0.14495968572278384}, "confidence": 0.11549479373450144, "choice": "hypatia"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 94.53812499850756, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.125, -3.03125, -0.66796875, -1.09375, -1.6484375, -1.5859375], "runtime_probabilities": [0.08455423265695572, 0.03416299447417259, 0.3630078136920929, 0.23713821172714233, 0.13617700338363647, 0.14495967328548431], "decide_seconds": 0.09002899900224293, "probabilities_f64_minus_runtime_max_abs": 2.3562624240636154e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0845542359698426, "einstein": 0.03416299956226512, "hypatia": 0.36300783725471714, "ibn_sina": 0.237138219141072, "sagan": 0.13617702234931925, "socrates": 0.14495968572278384}, "confidence": 0.11549479373450144, "seconds": 0.09456497499922989, "prompt_tokens": 380, "eval_tokens": 1, "server_latency_ms": 94.53812499850756, "prompt_eval_ms": null, "total_ms": 94.56497499922989, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:58Z"}
{"mode": "kev-choice", "prompt_sha256": "ef89dd562e7e2022614d3aeddd490894007adab3496f15b911b8f50b5a8695bf", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.012093397863395144, "einstein": 0.09814143696714911, "hypatia": 0.2763217430787911, "ibn_sina": 0.24672754535391742, "sagan": 0.18192548361811864, "socrates": 0.18479039311862863}, "confidence": 0.10481387404963893, "choice": "hypatia"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 94.29751700008637, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-4.125, -2.03125, -0.99609375, -1.109375, -1.4140625, -1.3984375], "runtime_probabilities": [0.012093398720026016, 0.09814143925905228, 0.27632173895835876, 0.24672754108905792, 0.18192549049854279, 0.18479038774967194], "decide_seconds": 0.08995649999997113, "probabilities_f64_minus_runtime_max_abs": 6.880424147093223e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.012093397863395144, "einstein": 0.09814143696714911, "hypatia": 0.2763217430787911, "ibn_sina": 0.24672754535391742, "sagan": 0.18192548361811864, "socrates": 0.18479039311862863}, "confidence": 0.10481387404963893, "seconds": 0.09432291600023746, "prompt_tokens": 349, "eval_tokens": 1, "server_latency_ms": 94.29751700008637, "prompt_eval_ms": null, "total_ms": 94.32291600023746, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:58Z"}
{"mode": "kev-choice", "prompt_sha256": "dca49448005572a62fd417c9eaf61d7d31ad351962f9ace12cbbf8888be7a516", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.020884612745877863, "einstein": 0.10606081495279988, "hypatia": 0.2827268976019075, "ibn_sina": 0.24853288121701142, "sagan": 0.17623620260996234, "socrates": 0.165558590872441}, "confidence": 0.09272772329744661, "choice": "hypatia"}}, "usage": {"input_tokens": 287, "output_tokens": 1}, "latency_ms": 90.98880199962878, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.203125, -1.578125, -0.59765625, -0.7265625, -1.0703125, -1.1328125], "runtime_probabilities": [0.020884612575173378, 0.10606081783771515, 0.282726913690567, 0.24853289127349854, 0.17623621225357056, 0.1655585914850235], "decide_seconds": 0.08736317900184076, "probabilities_f64_minus_runtime_max_abs": 1.6088659515745007e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.020884612745877863, "einstein": 0.10606081495279988, "hypatia": 0.2827268976019075, "ibn_sina": 0.24853288121701142, "sagan": 0.17623620260996234, "socrates": 0.165558590872441}, "confidence": 0.09272772329744661, "seconds": 0.09101231200111215, "prompt_tokens": 287, "eval_tokens": 1, "server_latency_ms": 90.98880199962878, "prompt_eval_ms": null, "total_ms": 91.01231200111215, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:58Z"}
{"mode": "kev-choice", "prompt_sha256": "78d241562930238f76945a3734169698bc708243442eeb4f19da867c434d4b2c", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.024593245560864754, "einstein": 0.09281393911939549, "hypatia": 0.3530233967706143, "ibn_sina": 0.196487110657648, "sagan": 0.12392286734763087, "socrates": 0.20915944054384658}, "confidence": 0.11535261873949876, "choice": "hypatia"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 103.93662199930986, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.4375, -2.109375, -0.7734375, -1.359375, -1.8203125, -1.296875], "runtime_probabilities": [0.02459324523806572, 0.09281393885612488, 0.35302338004112244, 0.19648709893226624, 0.12392286956310272, 0.20915944874286652], "decide_seconds": 0.0988091120016179, "probabilities_f64_minus_runtime_max_abs": 1.6729491847389255e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.024593245560864754, "einstein": 0.09281393911939549, "hypatia": 0.3530233967706143, "ibn_sina": 0.196487110657648, "sagan": 0.12392286734763087, "socrates": 0.20915944054384658}, "confidence": 0.11535261873949876, "seconds": 0.10396760200092103, "prompt_tokens": 407, "eval_tokens": 1, "server_latency_ms": 103.93662199930986, "prompt_eval_ms": null, "total_ms": 103.96760200092103, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:58Z"}
{"mode": "kev-choice", "prompt_sha256": "18ddc18998ab3fb1e833f98202c0e822e1cd8155ccbf0a8a97fcba7933b1f73c", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.009707111220391931, "einstein": 0.1845962914873085, "hypatia": 0.20353286333126058, "ibn_sina": 0.21623685490885328, "sagan": 0.22617210418781938, "socrates": 0.15975477486436637}, "confidence": 0.08400803956135583, "choice": "sagan"}}, "usage": {"input_tokens": 299, "output_tokens": 1}, "latency_ms": 91.69516700058011, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.546875, -0.6015625, -0.50390625, -0.443359375, -0.3984375, -0.74609375], "runtime_probabilities": [0.009707110933959484, 0.18459630012512207, 0.2035328894853592, 0.21623685956001282, 0.22617211937904358, 0.15975478291511536], "decide_seconds": 0.08785432600052445, "probabilities_f64_minus_runtime_max_abs": 2.6154098609065812e-08}}}, "choice": "sagan", "probs": {"darwin": 0.009707111220391931, "einstein": 0.1845962914873085, "hypatia": 0.20353286333126058, "ibn_sina": 0.21623685490885328, "sagan": 0.22617210418781938, "socrates": 0.15975477486436637}, "confidence": 0.08400803956135583, "seconds": 0.0917227069985529, "prompt_tokens": 299, "eval_tokens": 1, "server_latency_ms": 91.69516700058011, "prompt_eval_ms": null, "total_ms": 91.7227069985529, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:58Z"}
{"mode": "kev-choice", "prompt_sha256": "0cba123d72e8f7f1c3c7f88c01d5c38838fcf30e5686570d971c7add87b763a2", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.02789200303243803, "einstein": 0.10692097584284467, "hypatia": 0.25152911998843036, "ibn_sina": 0.2254692459576666, "sagan": 0.20053649859889625, "socrates": 0.18765215657972412}, "confidence": 0.07461036151103995, "choice": "hypatia"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 94.50403500159155, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.703125, -1.359375, -0.50390625, -0.61328125, -0.73046875, -0.796875], "runtime_probabilities": [0.02789200097322464, 0.10692097246646881, 0.25152912735939026, 0.22546924650669098, 0.20053650438785553, 0.18765215575695038], "decide_seconds": 0.09027445799802081, "probabilities_f64_minus_runtime_max_abs": 7.370959898000962e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.02789200303243803, "einstein": 0.10692097584284467, "hypatia": 0.25152911998843036, "ibn_sina": 0.2254692459576666, "sagan": 0.20053649859889625, "socrates": 0.18765215657972412}, "confidence": 0.07461036151103995, "seconds": 0.0945312850017217, "prompt_tokens": 346, "eval_tokens": 1, "server_latency_ms": 94.50403500159155, "prompt_eval_ms": null, "total_ms": 94.5312850017217, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:58Z"}
{"mode": "kev-choice", "prompt_sha256": "af92cf8e77853f7e17b471fab0386d943067a2e1b42c4599e1fd1b6432fd3b11", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.028198852231028276, "einstein": 0.2019522485397658, "hypatia": 0.1477514213326093, "ibn_sina": 0.21666313543325386, "sagan": 0.24937774025834403, "socrates": 0.15605660220499878}, "confidence": 0.06582945975933285, "choice": "sagan"}}, "usage": {"input_tokens": 384, "output_tokens": 1}, "latency_ms": 94.38596599648008, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.234375, -1.265625, -1.578125, -1.1953125, -1.0546875, -1.5234375], "runtime_probabilities": [0.0281988512724638, 0.2019522488117218, 0.14775142073631287, 0.21666312217712402, 0.24937772750854492, 0.15605659782886505], "decide_seconds": 0.0897846309999295, "probabilities_f64_minus_runtime_max_abs": 1.3256129832717178e-08}}}, "choice": "sagan", "probs": {"darwin": 0.028198852231028276, "einstein": 0.2019522485397658, "hypatia": 0.1477514213326093, "ibn_sina": 0.21666313543325386, "sagan": 0.24937774025834403, "socrates": 0.15605660220499878}, "confidence": 0.06582945975933285, "seconds": 0.09441343599974061, "prompt_tokens": 384, "eval_tokens": 1, "server_latency_ms": 94.38596599648008, "prompt_eval_ms": null, "total_ms": 94.41343599974061, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:58Z"}
{"mode": "kev-choice", "prompt_sha256": "fbe5ea900d8ce1bc3acd6791006648e0ced9aec40544e9afe4d501057d4ae8bd", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.05450761807834482, "einstein": 0.07747094579911525, "hypatia": 0.31428291962671484, "ibn_sina": 0.25302713176479086, "sagan": 0.1328161011317275, "socrates": 0.16789528359930678}, "confidence": 0.08695207448374076, "choice": "hypatia"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 94.3939560020226, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.203125, -1.8515625, -0.451171875, -0.66796875, -1.3125, -1.078125], "runtime_probabilities": [0.05450761690735817, 0.07747095078229904, 0.31428292393684387, 0.25302714109420776, 0.1328161060810089, 0.16789528727531433], "decide_seconds": 0.09011739900233806, "probabilities_f64_minus_runtime_max_abs": 9.329416905679011e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.05450761807834482, "einstein": 0.07747094579911525, "hypatia": 0.31428291962671484, "ibn_sina": 0.25302713176479086, "sagan": 0.1328161011317275, "socrates": 0.16789528359930678}, "confidence": 0.08695207448374076, "seconds": 0.0944214560004184, "prompt_tokens": 339, "eval_tokens": 1, "server_latency_ms": 94.3939560020226, "prompt_eval_ms": null, "total_ms": 94.4214560004184, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:58Z"}
{"mode": "kev-choice", "prompt_sha256": "7905abf1a7c38428613208ee4ff2e2dff37c5106352f2c0f16c09b8cf9fff9af", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.01578131313975046, "einstein": 0.1352688158923473, "hypatia": 0.28192442185247346, "ibn_sina": 0.27973046859567235, "sagan": 0.1241296429870496, "socrates": 0.16316533753270682}, "confidence": 0.10468429574450333, "choice": "hypatia"}}, "usage": {"input_tokens": 345, "output_tokens": 1}, "latency_ms": 94.36070599986124, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.625, -1.4765625, -0.7421875, -0.75, -1.5625, -1.2890625], "runtime_probabilities": [0.015781313180923462, 0.13526880741119385, 0.28192442655563354, 0.2797304391860962, 0.12412963807582855, 0.16316534578800201], "decide_seconds": 0.09013715900073294, "probabilities_f64_minus_runtime_max_abs": 2.9409576163264006e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.01578131313975046, "einstein": 0.1352688158923473, "hypatia": 0.28192442185247346, "ibn_sina": 0.27973046859567235, "sagan": 0.1241296429870496, "socrates": 0.16316533753270682}, "confidence": 0.10468429574450333, "seconds": 0.09438693599804537, "prompt_tokens": 345, "eval_tokens": 1, "server_latency_ms": 94.36070599986124, "prompt_eval_ms": null, "total_ms": 94.38693599804537, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:58Z"}
{"mode": "kev-choice", "prompt_sha256": "046c679c54add8987c0def6ffb019370532501b30ece1b193e34c4ac614f29e3", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.01846707363553385, "einstein": 0.31975396569475506, "hypatia": 0.07897365713468858, "ibn_sina": 0.18219031779247222, "sagan": 0.1684984991898511, "socrates": 0.23211648655269915}, "confidence": 0.11367601589261844, "choice": "einstein"}}, "usage": {"input_tokens": 358, "output_tokens": 1}, "latency_ms": 94.29629700025544, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.4375, -0.5859375, -1.984375, -1.1484375, -1.2265625, -0.90625], "runtime_probabilities": [0.018467074260115623, 0.3197539746761322, 0.07897365838289261, 0.18219031393527985, 0.16849851608276367, 0.2321164906024933], "decide_seconds": 0.09001629900012631, "probabilities_f64_minus_runtime_max_abs": 1.6892912568788887e-08}}}, "choice": "einstein", "probs": {"darwin": 0.01846707363553385, "einstein": 0.31975396569475506, "hypatia": 0.07897365713468858, "ibn_sina": 0.18219031779247222, "sagan": 0.1684984991898511, "socrates": 0.23211648655269915}, "confidence": 0.11367601589261844, "seconds": 0.09432176699920092, "prompt_tokens": 358, "eval_tokens": 1, "server_latency_ms": 94.29629700025544, "prompt_eval_ms": null, "total_ms": 94.32176699920092, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:58Z"}
{"mode": "kev-choice", "prompt_sha256": "1c86e1f79ad9ee96dfc6e80b1c4bfb37ccb503740730c4dcdc4de0bbe9576979", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.012478552611218333, "einstein": 0.7224188747146433, "hypatia": 0.01172251532798371, "ibn_sina": 0.07525530349995906, "sagan": 0.09738759874058321, "socrates": 0.08073715510561237}, "confidence": 0.4606447007115345, "choice": "einstein"}}, "usage": {"input_tokens": 402, "output_tokens": 1}, "latency_ms": 97.87236000192934, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.234375, 0.82421875, -3.296875, -1.4375, -1.1796875, -1.3671875], "runtime_probabilities": [0.012478552758693695, 0.7224189043045044, 0.011722516268491745, 0.07525530457496643, 0.09738760441541672, 0.0807371586561203], "decide_seconds": 0.09309941600076854, "probabilities_f64_minus_runtime_max_abs": 2.9589861116896543e-08}}}, "choice": "einstein", "probs": {"darwin": 0.012478552611218333, "einstein": 0.7224188747146433, "hypatia": 0.01172251532798371, "ibn_sina": 0.07525530349995906, "sagan": 0.09738759874058321, "socrates": 0.08073715510561237}, "confidence": 0.4606447007115345, "seconds": 0.09789981900030398, "prompt_tokens": 402, "eval_tokens": 1, "server_latency_ms": 97.87236000192934, "prompt_eval_ms": null, "total_ms": 97.89981900030398, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:59Z"}
{"mode": "kev-choice", "prompt_sha256": "08395ee9a394a652ac502ba3783881f851a097c91c2d8c801bc9dd19465066fa", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.035362099903338676, "einstein": 0.5614558826339887, "hypatia": 0.022831487693986522, "ibn_sina": 0.11415025677351916, "sagan": 0.11962844197653626, "socrates": 0.1465718310186306}, "confidence": 0.26789034238155607, "choice": "einstein"}}, "usage": {"input_tokens": 305, "output_tokens": 1}, "latency_ms": 91.38279900071211, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.8125, -0.047607421875, -3.25, -1.640625, -1.59375, -1.390625], "runtime_probabilities": [0.03536210209131241, 0.5614559054374695, 0.02283148653805256, 0.11415025591850281, 0.1196284368634224, 0.1465718299150467], "decide_seconds": 0.08755595899856417, "probabilities_f64_minus_runtime_max_abs": 2.2803480770328122e-08}}}, "choice": "einstein", "probs": {"darwin": 0.035362099903338676, "einstein": 0.5614558826339887, "hypatia": 0.022831487693986522, "ibn_sina": 0.11415025677351916, "sagan": 0.11962844197653626, "socrates": 0.1465718310186306}, "confidence": 0.26789034238155607, "seconds": 0.0914068079982826, "prompt_tokens": 305, "eval_tokens": 1, "server_latency_ms": 91.38279900071211, "prompt_eval_ms": null, "total_ms": 91.4068079982826, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:59Z"}
{"mode": "kev-choice", "prompt_sha256": "06550a0c070754bfa382273603bae053348544611ea06fc60c3f8f78f6f0e0b7", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0063022427433428825, "einstein": 0.4833540031908388, "hypatia": 0.04478364747434313, "ibn_sina": 0.12365161575587702, "sagan": 0.24783930217920144, "socrates": 0.09406918865639677}, "confidence": 0.24712078725338227, "choice": "einstein"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 90.67883500028984, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.421875, 0.91796875, -1.4609375, -0.4453125, 0.25, -0.71875], "runtime_probabilities": [0.006302242167294025, 0.48335397243499756, 0.04478364437818527, 0.12365161627531052, 0.24783927202224731, 0.09406919032335281], "decide_seconds": 0.08713803099817596, "probabilities_f64_minus_runtime_max_abs": 3.075584126266406e-08}}}, "choice": "einstein", "probs": {"darwin": 0.0063022427433428825, "einstein": 0.4833540031908388, "hypatia": 0.04478364747434313, "ibn_sina": 0.12365161575587702, "sagan": 0.24783930217920144, "socrates": 0.09406918865639677}, "confidence": 0.24712078725338227, "seconds": 0.09070964499915135, "prompt_tokens": 267, "eval_tokens": 1, "server_latency_ms": 90.67883500028984, "prompt_eval_ms": null, "total_ms": 90.70964499915135, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:59Z"}
{"mode": "kev-choice", "prompt_sha256": "8d3b9b5c684b9c4f71fe2c80ef2a0abb1bf762cb9eca32cc2797975293e622db", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0298516968594907, "einstein": 0.25787930991265096, "hypatia": 0.08372115424381067, "ibn_sina": 0.25190554213262917, "sagan": 0.1678052673924492, "socrates": 0.20883702945896926}, "confidence": 0.08699964175135921, "choice": "einstein"}}, "usage": {"input_tokens": 294, "output_tokens": 1}, "latency_ms": 90.88759299993399, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.0625, -0.90625, -2.03125, -0.9296875, -1.3359375, -1.1171875], "runtime_probabilities": [0.029851699247956276, 0.2578793168067932, 0.08372115343809128, 0.25190556049346924, 0.1678052693605423, 0.20883704721927643], "decide_seconds": 0.08716097200158401, "probabilities_f64_minus_runtime_max_abs": 1.836084007056371e-08}}}, "choice": "einstein", "probs": {"darwin": 0.0298516968594907, "einstein": 0.25787930991265096, "hypatia": 0.08372115424381067, "ibn_sina": 0.25190554213262917, "sagan": 0.1678052673924492, "socrates": 0.20883702945896926}, "confidence": 0.08699964175135921, "seconds": 0.0909121229997254, "prompt_tokens": 294, "eval_tokens": 1, "server_latency_ms": 90.88759299993399, "prompt_eval_ms": null, "total_ms": 90.9121229997254, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:59Z"}
{"mode": "kev-choice", "prompt_sha256": "ef45da2316173e7abeea5d9f50b805e7e6101970e193e303448a09b24841b95c", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.014206015535044036, "einstein": 0.29554519051253547, "hypatia": 0.11528594311834317, "ibn_sina": 0.17104843365212316, "sagan": 0.20471817027785322, "socrates": 0.19919624690410098}, "confidence": 0.09704223895333619, "choice": "einstein"}}, "usage": {"input_tokens": 387, "output_tokens": 1}, "latency_ms": 98.44988500117324, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.484375, -0.44921875, -1.390625, -0.99609375, -0.81640625, -0.84375], "runtime_probabilities": [0.01420601550489664, 0.29554519057273865, 0.11528594046831131, 0.17104843258857727, 0.20471817255020142, 0.1991962492465973], "decide_seconds": 0.09361414200247964, "probabilities_f64_minus_runtime_max_abs": 2.650031860618185e-09}}}, "choice": "einstein", "probs": {"darwin": 0.014206015535044036, "einstein": 0.29554519051253547, "hypatia": 0.11528594311834317, "ibn_sina": 0.17104843365212316, "sagan": 0.20471817027785322, "socrates": 0.19919624690410098}, "confidence": 0.09704223895333619, "seconds": 0.0984784549982578, "prompt_tokens": 387, "eval_tokens": 1, "server_latency_ms": 98.44988500117324, "prompt_eval_ms": null, "total_ms": 98.4784549982578, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:59Z"}
{"mode": "kev-choice", "prompt_sha256": "70c61a24e7389825115e809e90590b0de304acdcec9da8139878d698c530832c", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.005989329111324775, "einstein": 0.7905998051517978, "hypatia": 0.007338279425157467, "ibn_sina": 0.053800991932586494, "sagan": 0.07127473411678524, "socrates": 0.07099686026234829}, "confidence": 0.5614603774121277, "choice": "einstein"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 95.02455100300722, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.421875, 1.4609375, -3.21875, -1.2265625, -0.9453125, -0.94921875], "runtime_probabilities": [0.00598932895809412, 0.7905997633934021, 0.007338279392570257, 0.05380098894238472, 0.07127472758293152, 0.07099685817956924], "decide_seconds": 0.09064319499884732, "probabilities_f64_minus_runtime_max_abs": 4.175839574127593e-08}}}, "choice": "einstein", "probs": {"darwin": 0.005989329111324775, "einstein": 0.7905998051517978, "hypatia": 0.007338279425157467, "ibn_sina": 0.053800991932586494, "sagan": 0.07127473411678524, "socrates": 0.07099686026234829}, "confidence": 0.5614603774121277, "seconds": 0.09505145999719389, "prompt_tokens": 343, "eval_tokens": 1, "server_latency_ms": 95.02455100300722, "prompt_eval_ms": null, "total_ms": 95.05145999719389, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:59Z"}
{"mode": "kev-choice", "prompt_sha256": "36e36fd875dcb6fcd30f8ad4a732fc84604c5605fd51588de8cdf3fd37bcb156", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0035389992211713093, "einstein": 0.03154275670883777, "hypatia": 0.5110667619574781, "ibn_sina": 0.1569357002314884, "sagan": 0.13099669662668675, "socrates": 0.16591908525433766}, "confidence": 0.259397478979539, "choice": "hypatia"}}, "usage": {"input_tokens": 278, "output_tokens": 1}, "latency_ms": 93.91393000259995, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-4.0, -1.8125, 0.97265625, -0.2080078125, -0.388671875, -0.15234375], "runtime_probabilities": [0.003538999008014798, 0.03154275566339493, 0.5110667943954468, 0.1569357067346573, 0.1309966892004013, 0.16591908037662506], "decide_seconds": 0.09039909599960083, "probabilities_f64_minus_runtime_max_abs": 3.243796864449422e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0035389992211713093, "einstein": 0.03154275670883777, "hypatia": 0.5110667619574781, "ibn_sina": 0.1569357002314884, "sagan": 0.13099669662668675, "socrates": 0.16591908525433766}, "confidence": 0.259397478979539, "seconds": 0.09393821000048774, "prompt_tokens": 278, "eval_tokens": 1, "server_latency_ms": 93.91393000259995, "prompt_eval_ms": null, "total_ms": 93.93821000048774, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:59Z"}
{"mode": "kev-choice", "prompt_sha256": "53410880d12779c0c4c4e4becb3d6c57efa5cca73c859c565c763499a1fba816", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.025666332646564937, "einstein": 0.0784426370404459, "hypatia": 0.41342536166204824, "ibn_sina": 0.19113741494314854, "sagan": 0.11593070238123522, "socrates": 0.17539755132655707}, "confidence": 0.14595334107621905, "choice": "hypatia"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 97.56870199998957, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.09375, -1.9765625, -0.314453125, -1.0859375, -1.5859375, -1.171875], "runtime_probabilities": [0.025666331872344017, 0.07844263315200806, 0.41342535614967346, 0.1911374032497406, 0.11593069136142731, 0.17539756000041962], "decide_seconds": 0.09280117799789878, "probabilities_f64_minus_runtime_max_abs": 1.169340793638618e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.025666332646564937, "einstein": 0.0784426370404459, "hypatia": 0.41342536166204824, "ibn_sina": 0.19113741494314854, "sagan": 0.11593070238123522, "socrates": 0.17539755132655707}, "confidence": 0.14595334107621905, "seconds": 0.0975955609974335, "prompt_tokens": 395, "eval_tokens": 1, "server_latency_ms": 97.56870199998957, "prompt_eval_ms": null, "total_ms": 97.5955609974335, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:59Z"}
{"mode": "kev-choice", "prompt_sha256": "60ca01400728b39379b62e800f8729bb21f28627f8ef74fdb91406ffc80d039b", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.01563164230534739, "einstein": 0.5036554622584912, "hypatia": 0.06843331126189717, "ibn_sina": 0.17204132555895013, "sagan": 0.10935612002156711, "socrates": 0.13088213859374703}, "confidence": 0.2158915622112053, "choice": "einstein"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 94.0510379987245, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.46875, 0.00384521484375, -1.9921875, -1.0703125, -1.5234375, -1.34375], "runtime_probabilities": [0.01563164032995701, 0.5036554336547852, 0.06843330711126328, 0.17204132676124573, 0.1093561127781868, 0.130882129073143], "decide_seconds": 0.09006145900275442, "probabilities_f64_minus_runtime_max_abs": 2.860370607038476e-08}}}, "choice": "einstein", "probs": {"darwin": 0.01563164230534739, "einstein": 0.5036554622584912, "hypatia": 0.06843331126189717, "ibn_sina": 0.17204132555895013, "sagan": 0.10935612002156711, "socrates": 0.13088213859374703}, "confidence": 0.2158915622112053, "seconds": 0.09407569800168858, "prompt_tokens": 323, "eval_tokens": 1, "server_latency_ms": 94.0510379987245, "prompt_eval_ms": null, "total_ms": 94.07569800168858, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:59Z"}
{"mode": "kev-choice", "prompt_sha256": "cea561710397d08713919835eaa6badd506abefe21e9793915736cbcf571db70", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0569583728538533, "einstein": 0.06870491229411668, "hypatia": 0.3214331049155559, "ibn_sina": 0.2351654832170663, "sagan": 0.13243218462124628, "socrates": 0.18530594209816154}, "confidence": 0.08887190114333798, "choice": "hypatia"}}, "usage": {"input_tokens": 274, "output_tokens": 1}, "latency_ms": 90.61858499990194, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.21875, -2.03125, -0.48828125, -0.80078125, -1.375, -1.0390625], "runtime_probabilities": [0.05695836618542671, 0.06870491057634354, 0.32143309712409973, 0.23516547679901123, 0.13243217766284943, 0.18530595302581787], "decide_seconds": 0.08709790200009593, "probabilities_f64_minus_runtime_max_abs": 1.0927656335368852e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0569583728538533, "einstein": 0.06870491229411668, "hypatia": 0.3214331049155559, "ibn_sina": 0.2351654832170663, "sagan": 0.13243218462124628, "socrates": 0.18530594209816154}, "confidence": 0.08887190114333798, "seconds": 0.09064167500037001, "prompt_tokens": 274, "eval_tokens": 1, "server_latency_ms": 90.61858499990194, "prompt_eval_ms": null, "total_ms": 90.64167500037001, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:49:59Z"}
{"mode": "kev-choice", "prompt_sha256": "a92667bd1952495e79468ae067425395d5511079bdf1f066abafa0182c57620e", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.015689127281002507, "einstein": 0.3302476879839176, "hypatia": 0.10975827443211343, "ibn_sina": 0.16605896607020826, "sagan": 0.14769590183011014, "socrates": 0.23055004240264798}, "confidence": 0.11121482582421216, "choice": "einstein"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 91.37920899956953, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.8125, -0.765625, -1.8671875, -1.453125, -1.5703125, -1.125], "runtime_probabilities": [0.015689127147197723, 0.330247700214386, 0.10975827276706696, 0.1660589724779129, 0.1476958990097046, 0.23055005073547363], "decide_seconds": 0.08760066800095956, "probabilities_f64_minus_runtime_max_abs": 1.2230468360296953e-08}}}, "choice": "einstein", "probs": {"darwin": 0.015689127281002507, "einstein": 0.3302476879839176, "hypatia": 0.10975827443211343, "ibn_sina": 0.16605896607020826, "sagan": 0.14769590183011014, "socrates": 0.23055004240264798}, "confidence": 0.11121482582421216, "seconds": 0.0914026489990647, "prompt_tokens": 298, "eval_tokens": 1, "server_latency_ms": 91.37920899956953, "prompt_eval_ms": null, "total_ms": 91.4026489990647, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:00Z"}
{"mode": "kev-choice", "prompt_sha256": "6dae1f58ef7ea31d6acf53cc5f4fc8a8254827e1ec602b97859fd641edded372", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.057789579907591045, "einstein": 0.0542882862596942, "hypatia": 0.3609862002005115, "ibn_sina": 0.1939782386784125, "sagan": 0.13125224170865207, "socrates": 0.20170545324513878}, "confidence": 0.10796808988968987, "choice": "hypatia"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 94.20665800280403, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.25, -2.3125, -0.41796875, -1.0390625, -1.4296875, -1.0], "runtime_probabilities": [0.057789579033851624, 0.054288286715745926, 0.36098620295524597, 0.1939782351255417, 0.1312522441148758, 0.2017054408788681], "decide_seconds": 0.0897766509988287, "probabilities_f64_minus_runtime_max_abs": 1.2366270674135649e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.057789579907591045, "einstein": 0.0542882862596942, "hypatia": 0.3609862002005115, "ibn_sina": 0.1939782386784125, "sagan": 0.13125224170865207, "socrates": 0.20170545324513878}, "confidence": 0.10796808988968987, "seconds": 0.09423248799794237, "prompt_tokens": 368, "eval_tokens": 1, "server_latency_ms": 94.20665800280403, "prompt_eval_ms": null, "total_ms": 94.23248799794237, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:00Z"}
{"mode": "kev-choice", "prompt_sha256": "c5e40b39c14beb9a0b05e0922a154ec9943477a8ebdd3b239151420d3872133c", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.28876275156016984, "einstein": 0.05947328957826725, "hypatia": 0.22097007765135804, "ibn_sina": 0.13298213516183474, "sagan": 0.1035665909985443, "socrates": 0.19424515504982579}, "confidence": 0.06149593739353598, "choice": "darwin"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 91.20583100229851, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-0.255859375, -1.8359375, -0.5234375, -1.03125, -1.28125, -0.65234375], "runtime_probabilities": [0.2887627184391022, 0.05947328358888626, 0.2209700495004654, 0.13298211991786957, 0.10356658697128296, 0.1942451447248459], "decide_seconds": 0.08736372999919695, "probabilities_f64_minus_runtime_max_abs": 3.312106766628631e-08}}}, "choice": "darwin", "probs": {"darwin": 0.28876275156016984, "einstein": 0.05947328957826725, "hypatia": 0.22097007765135804, "ibn_sina": 0.13298213516183474, "sagan": 0.1035665909985443, "socrates": 0.19424515504982579}, "confidence": 0.06149593739353598, "seconds": 0.09123022099811351, "prompt_tokens": 297, "eval_tokens": 1, "server_latency_ms": 91.20583100229851, "prompt_eval_ms": null, "total_ms": 91.23022099811351, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:00Z"}
{"mode": "kev-choice", "prompt_sha256": "79fb4c421b342027371a354bfc724fa64f5db5d74a395e1f0134e383b6932a60", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.013626863707350615, "einstein": 0.09458918225695095, "hypatia": 0.3748388656608887, "ibn_sina": 0.19484402506913584, "sagan": 0.13762644809961994, "socrates": 0.18447461520605388}, "confidence": 0.1333433750844305, "choice": "hypatia"}}, "usage": {"input_tokens": 388, "output_tokens": 1}, "latency_ms": 97.41114300049958, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.421875, -1.484375, -0.107421875, -0.76171875, -1.109375, -0.81640625], "runtime_probabilities": [0.01362686324864626, 0.09458918124437332, 0.37483885884284973, 0.19484403729438782, 0.13762645423412323, 0.1844746172428131], "decide_seconds": 0.0927203789979103, "probabilities_f64_minus_runtime_max_abs": 1.2225251977415752e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.013626863707350615, "einstein": 0.09458918225695095, "hypatia": 0.3748388656608887, "ibn_sina": 0.19484402506913584, "sagan": 0.13762644809961994, "socrates": 0.18447461520605388}, "confidence": 0.1333433750844305, "seconds": 0.09743841200179304, "prompt_tokens": 388, "eval_tokens": 1, "server_latency_ms": 97.41114300049958, "prompt_eval_ms": null, "total_ms": 97.43841200179304, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:00Z"}
{"mode": "kev-choice", "prompt_sha256": "1171383a350272ae6dd50d3be1dda29c6229f3879efb1331a0015d33d4fa3568", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.01507639657284136, "einstein": 0.1114003400464506, "hypatia": 0.2730390308112027, "ibn_sina": 0.23217824636506867, "sagan": 0.13972732990786266, "socrates": 0.22857865629657403}, "confidence": 0.09946133233178267, "choice": "hypatia"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 94.05515800244757, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.328125, -1.328125, -0.431640625, -0.59375, -1.1015625, -0.609375], "runtime_probabilities": [0.015076395124197006, 0.11140033602714539, 0.2730390131473541, 0.2321782410144806, 0.13972732424736023, 0.22857865691184998], "decide_seconds": 0.08986463100154651, "probabilities_f64_minus_runtime_max_abs": 1.7663848550508732e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.01507639657284136, "einstein": 0.1114003400464506, "hypatia": 0.2730390308112027, "ibn_sina": 0.23217824636506867, "sagan": 0.13972732990786266, "socrates": 0.22857865629657403}, "confidence": 0.09946133233178267, "seconds": 0.09408156800054712, "prompt_tokens": 339, "eval_tokens": 1, "server_latency_ms": 94.05515800244757, "prompt_eval_ms": null, "total_ms": 94.08156800054712, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:00Z"}
{"mode": "kev-choice", "prompt_sha256": "550dcbbf4c9113c575eb00cf7dd4c978283e0d887b4fa383e793b9d3768433b0", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0023829129626818874, "einstein": 0.9173129346369097, "hypatia": 0.0017708305488850005, "ibn_sina": 0.02115587031376002, "sagan": 0.025618728683065247, "socrates": 0.031758722854698176}, "confidence": 0.7824549962815928, "choice": "einstein"}}, "usage": {"input_tokens": 275, "output_tokens": 1}, "latency_ms": 90.96819199839956, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.15625, 2.796875, -3.453125, -0.97265625, -0.78125, -0.56640625], "runtime_probabilities": [0.0023829129058867693, 0.9173129796981812, 0.00177083071321249, 0.02115587145090103, 0.02561873011291027, 0.03175872191786766], "decide_seconds": 0.08724341099878075, "probabilities_f64_minus_runtime_max_abs": 4.5061271491952937e-08}}}, "choice": "einstein", "probs": {"darwin": 0.0023829129626818874, "einstein": 0.9173129346369097, "hypatia": 0.0017708305488850005, "ibn_sina": 0.02115587031376002, "sagan": 0.025618728683065247, "socrates": 0.031758722854698176}, "confidence": 0.7824549962815928, "seconds": 0.09099219199924846, "prompt_tokens": 275, "eval_tokens": 1, "server_latency_ms": 90.96819199839956, "prompt_eval_ms": null, "total_ms": 90.99219199924846, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:00Z"}
{"mode": "kev-choice", "prompt_sha256": "a7501a3b3149ec22eb93de5cfbdf2becc70db679c219e6f88ab7072e69212da2", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.050560728954299014, "einstein": 0.21622081287603545, "hypatia": 0.1113009359325282, "ibn_sina": 0.17647428422977457, "sagan": 0.17104474130592207, "socrates": 0.27439849670144073}, "confidence": 0.05713467063100586, "choice": "socrates"}}, "usage": {"input_tokens": 349, "output_tokens": 1}, "latency_ms": 95.28364899961161, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.390625, -0.9375, -1.6015625, -1.140625, -1.171875, -0.69921875], "runtime_probabilities": [0.05056072399020195, 0.21622081100940704, 0.11130094528198242, 0.17647428810596466, 0.1710447371006012, 0.27439847588539124], "decide_seconds": 0.09098234300108743, "probabilities_f64_minus_runtime_max_abs": 2.0816049495220312e-08}}}, "choice": "socrates", "probs": {"darwin": 0.050560728954299014, "einstein": 0.21622081287603545, "hypatia": 0.1113009359325282, "ibn_sina": 0.17647428422977457, "sagan": 0.17104474130592207, "socrates": 0.27439849670144073}, "confidence": 0.05713467063100586, "seconds": 0.09530992900181445, "prompt_tokens": 349, "eval_tokens": 1, "server_latency_ms": 95.28364899961161, "prompt_eval_ms": null, "total_ms": 95.30992900181445, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:00Z"}
{"mode": "kev-choice", "prompt_sha256": "51091ba903351f76f1c9074582ea4abcadfbf07b73645803c2453268f86cf0fe", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.08874685127561345, "einstein": 0.09823380957616809, "hypatia": 0.20314356567258265, "ibn_sina": 0.1820967153841103, "sagan": 0.18935062387021565, "socrates": 0.2384284342213098}, "confidence": 0.03237371524084076, "choice": "socrates"}}, "usage": {"input_tokens": 364, "output_tokens": 1}, "latency_ms": 94.76755300056539, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-1.875, -1.7734375, -1.046875, -1.15625, -1.1171875, -0.88671875], "runtime_probabilities": [0.08874684572219849, 0.09823380410671234, 0.20314353704452515, 0.1820967048406601, 0.18935061991214752, 0.23842841386795044], "decide_seconds": 0.09026921699842205, "probabilities_f64_minus_runtime_max_abs": 2.862805750791786e-08}}}, "choice": "socrates", "probs": {"darwin": 0.08874685127561345, "einstein": 0.09823380957616809, "hypatia": 0.20314356567258265, "ibn_sina": 0.1820967153841103, "sagan": 0.18935062387021565, "socrates": 0.2384284342213098}, "confidence": 0.03237371524084076, "seconds": 0.09479494300103397, "prompt_tokens": 364, "eval_tokens": 1, "server_latency_ms": 94.76755300056539, "prompt_eval_ms": null, "total_ms": 94.79494300103397, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:00Z"}
{"mode": "kev-choice", "prompt_sha256": "ef8ed0aee48a818fd4b6f519a6d69a1ee3cdac3aca20b9c6b8fec197d1a29a2f", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.021466132105262076, "einstein": 0.08828223888112205, "hypatia": 0.30070354518635595, "ibn_sina": 0.2303323618132896, "sagan": 0.13943104698330247, "socrates": 0.21978467503066787}, "confidence": 0.1048153303417787, "choice": "hypatia"}}, "usage": {"input_tokens": 329, "output_tokens": 1}, "latency_ms": 94.31949599820655, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.78125, -1.3671875, -0.1416015625, -0.408203125, -0.91015625, -0.455078125], "runtime_probabilities": [0.021466132253408432, 0.08828224241733551, 0.3007035553455353, 0.2303323894739151, 0.13943105936050415, 0.2197846919298172], "decide_seconds": 0.0901807580012246, "probabilities_f64_minus_runtime_max_abs": 2.7660625512293535e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.021466132105262076, "einstein": 0.08828223888112205, "hypatia": 0.30070354518635595, "ibn_sina": 0.2303323618132896, "sagan": 0.13943104698330247, "socrates": 0.21978467503066787}, "confidence": 0.1048153303417787, "seconds": 0.09434730599969043, "prompt_tokens": 329, "eval_tokens": 1, "server_latency_ms": 94.31949599820655, "prompt_eval_ms": null, "total_ms": 94.34730599969043, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:00Z"}
{"mode": "kev-choice", "prompt_sha256": "518f70391848f4160de4d2ca4b3dfa9256f61398a04ebe65ce467d3f4df3d9ec", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.49088356309164033, "einstein": 0.02925043534896108, "hypatia": 0.1290589733039811, "ibn_sina": 0.08332673593601206, "sagan": 0.11659506598602341, "socrates": 0.15088522633338194}, "confidence": 0.18525001909502636, "choice": "darwin"}}, "usage": {"input_tokens": 407, "output_tokens": 1}, "latency_ms": 97.8318589986884, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [0.3828125, -2.4375, -0.953125, -1.390625, -1.0546875, -0.796875], "runtime_probabilities": [0.49088355898857117, 0.029250435531139374, 0.12905897200107574, 0.08332673460245132, 0.1165950670838356, 0.1508852243423462], "decide_seconds": 0.09298238699921058, "probabilities_f64_minus_runtime_max_abs": 4.103069162209039e-09}}}, "choice": "darwin", "probs": {"darwin": 0.49088356309164033, "einstein": 0.02925043534896108, "hypatia": 0.1290589733039811, "ibn_sina": 0.08332673593601206, "sagan": 0.11659506598602341, "socrates": 0.15088522633338194}, "confidence": 0.18525001909502636, "seconds": 0.09786122899822658, "prompt_tokens": 407, "eval_tokens": 1, "server_latency_ms": 97.8318589986884, "prompt_eval_ms": null, "total_ms": 97.86122899822658, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:00Z"}
{"mode": "kev-choice", "prompt_sha256": "2c8f8199cffc7ca570c61b59355b5cb2b0d15104ad1e57f3fd7692a8b2ec257e", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.00890854754975704, "einstein": 0.007619884519780672, "hypatia": 0.70769358799639, "ibn_sina": 0.08737544903397597, "sagan": 0.13297168551588784, "socrates": 0.05543084538420847}, "confidence": 0.46114099552657517, "choice": "hypatia"}}, "usage": {"input_tokens": 272, "output_tokens": 1}, "latency_ms": 90.45183600028395, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.453125, -2.609375, 1.921875, -0.169921875, 0.25, -0.625], "runtime_probabilities": [0.008908547461032867, 0.007619884796440601, 0.7076935768127441, 0.08737543970346451, 0.13297167420387268, 0.05543084442615509], "decide_seconds": 0.08695143400109373, "probabilities_f64_minus_runtime_max_abs": 1.1312015157916022e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.00890854754975704, "einstein": 0.007619884519780672, "hypatia": 0.70769358799639, "ibn_sina": 0.08737544903397597, "sagan": 0.13297168551588784, "socrates": 0.05543084538420847}, "confidence": 0.46114099552657517, "seconds": 0.09047549599927152, "prompt_tokens": 272, "eval_tokens": 1, "server_latency_ms": 90.45183600028395, "prompt_eval_ms": null, "total_ms": 90.47549599927152, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:00Z"}
{"mode": "kev-choice", "prompt_sha256": "49cf9584bdd714487fe44d03d0cf094bffb7ddabf39f43dc80799d3de073d3f7", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.04211273085463801, "einstein": 0.04771997582235266, "hypatia": 0.3410902024349675, "ibn_sina": 0.22500696124063818, "sagan": 0.16397682232879338, "socrates": 0.18009330731861026}, "confidence": 0.11467902231822569, "choice": "hypatia"}}, "usage": {"input_tokens": 326, "output_tokens": 1}, "latency_ms": 93.93883899974753, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.546875, -2.421875, -0.455078125, -0.87109375, -1.1875, -1.09375], "runtime_probabilities": [0.04211272671818733, 0.04771997034549713, 0.3410901725292206, 0.22500695288181305, 0.16397681832313538, 0.18009328842163086], "decide_seconds": 0.09001239000281203, "probabilities_f64_minus_runtime_max_abs": 2.990574693439996e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.04211273085463801, "einstein": 0.04771997582235266, "hypatia": 0.3410902024349675, "ibn_sina": 0.22500696124063818, "sagan": 0.16397682232879338, "socrates": 0.18009330731861026}, "confidence": 0.11467902231822569, "seconds": 0.09396280899818521, "prompt_tokens": 326, "eval_tokens": 1, "server_latency_ms": 93.93883899974753, "prompt_eval_ms": null, "total_ms": 93.96280899818521, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:01Z"}
{"mode": "kev-choice", "prompt_sha256": "d445b0eba928e2f123e2a696fc3349b65c89382bdce54490773f3593ea646bbd", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.019723330715322347, "einstein": 0.05319634735680152, "hypatia": 0.45419244018201443, "ibn_sina": 0.16131592918610302, "sagan": 0.15213541087483165, "socrates": 0.15943654168492694}, "confidence": 0.1821014388545179, "choice": "hypatia"}}, "usage": {"input_tokens": 303, "output_tokens": 1}, "latency_ms": 90.96174200021778, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.796875, -1.8046875, 0.33984375, -0.6953125, -0.75390625, -0.70703125], "runtime_probabilities": [0.019723327830433846, 0.053196344524621964, 0.4541924297809601, 0.1613159328699112, 0.15213540196418762, 0.15943653881549835], "decide_seconds": 0.08723556100085261, "probabilities_f64_minus_runtime_max_abs": 1.0401054351127925e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.019723330715322347, "einstein": 0.05319634735680152, "hypatia": 0.45419244018201443, "ibn_sina": 0.16131592918610302, "sagan": 0.15213541087483165, "socrates": 0.15943654168492694}, "confidence": 0.1821014388545179, "seconds": 0.09098608199929004, "prompt_tokens": 303, "eval_tokens": 1, "server_latency_ms": 90.96174200021778, "prompt_eval_ms": null, "total_ms": 90.98608199929004, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:01Z"}
{"mode": "kev-choice", "prompt_sha256": "44eede8e8a6cd0f0ec8f329973c8450624acfaaf39b0e500af444335f60a264d", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.004240407661253687, "einstein": 0.12732705602867958, "hypatia": 0.3728670460306549, "ibn_sina": 0.18207615545893624, "sagan": 0.23061580893497807, "socrates": 0.08287352588549747}, "confidence": 0.15821662306205808, "choice": "hypatia"}}, "usage": {"input_tokens": 273, "output_tokens": 1}, "latency_ms": 90.31720700295409, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.4375, -0.035400390625, 1.0390625, 0.322265625, 0.55859375, -0.46484375], "runtime_probabilities": [0.004240407608449459, 0.1273270547389984, 0.37286704778671265, 0.18207615613937378, 0.23061580955982208, 0.08287353068590164], "decide_seconds": 0.08676584400018328, "probabilities_f64_minus_runtime_max_abs": 4.800404171390049e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.004240407661253687, "einstein": 0.12732705602867958, "hypatia": 0.3728670460306549, "ibn_sina": 0.18207615545893624, "sagan": 0.23061580893497807, "socrates": 0.08287352588549747}, "confidence": 0.15821662306205808, "seconds": 0.09034004700151854, "prompt_tokens": 273, "eval_tokens": 1, "server_latency_ms": 90.31720700295409, "prompt_eval_ms": null, "total_ms": 90.34004700151854, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:01Z"}
{"mode": "kev-choice", "prompt_sha256": "a66cb47a052715ec47cff3da36e4d6f8643ac70737abfd23a65b0300a9c4f131", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0346582747696751, "einstein": 0.11187828141145167, "hypatia": 0.2946161215241564, "ibn_sina": 0.22911140780475503, "sagan": 0.21565103169861494, "socrates": 0.11408488279134685}, "confidence": 0.08597225108708229, "choice": "hypatia"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 91.14301099907607, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.15625, -0.984375, -0.01611328125, -0.267578125, -0.328125, -0.96484375], "runtime_probabilities": [0.034658271819353104, 0.11187827587127686, 0.29461610317230225, 0.22911138832569122, 0.2156510353088379, 0.1140848845243454], "decide_seconds": 0.08728879000045708, "probabilities_f64_minus_runtime_max_abs": 1.9479063806393526e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0346582747696751, "einstein": 0.11187828141145167, "hypatia": 0.2946161215241564, "ibn_sina": 0.22911140780475503, "sagan": 0.21565103169861494, "socrates": 0.11408488279134685}, "confidence": 0.08597225108708229, "seconds": 0.09116688100039028, "prompt_tokens": 309, "eval_tokens": 1, "server_latency_ms": 91.14301099907607, "prompt_eval_ms": null, "total_ms": 91.16688100039028, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:01Z"}
{"mode": "kev-choice", "prompt_sha256": "2bfa4bdd272a13a8e0a461cd9394d00a24d529a16ca98d344a304f473afee10d", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0022141081009880935, "einstein": 0.012941958877645033, "hypatia": 0.7821426643965369, "ibn_sina": 0.0553457382163534, "sagan": 0.06807642078336403, "socrates": 0.07927910962511249}, "confidence": 0.55013619918063, "choice": "hypatia"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 91.08878199913306, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-4.28125, -2.515625, 1.5859375, -1.0625, -0.85546875, -0.703125], "runtime_probabilities": [0.0022141083609312773, 0.01294195931404829, 0.782142698764801, 0.05534573644399643, 0.06807641685009003, 0.07927911728620529], "decide_seconds": 0.08721725099894684, "probabilities_f64_minus_runtime_max_abs": 3.4368264123507686e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0022141081009880935, "einstein": 0.012941958877645033, "hypatia": 0.7821426643965369, "ibn_sina": 0.0553457382163534, "sagan": 0.06807642078336403, "socrates": 0.07927910962511249}, "confidence": 0.55013619918063, "seconds": 0.0911125010024989, "prompt_tokens": 311, "eval_tokens": 1, "server_latency_ms": 91.08878199913306, "prompt_eval_ms": null, "total_ms": 91.1125010024989, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:01Z"}
{"mode": "kev-choice", "prompt_sha256": "a316eee1d777f5acec746f93cc5f922a09147773e8466eeda1d6cd77d962c38e", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.003365387612049128, "einstein": 0.014391937382428326, "hypatia": 0.6826915491513027, "ibn_sina": 0.09203205090481084, "sagan": 0.09607274316021788, "socrates": 0.11144633178919115}, "confidence": 0.42517543281735126, "choice": "hypatia"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 94.03084800214856, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-4.0, -2.546875, 1.3125, -0.69140625, -0.6484375, -0.5], "runtime_probabilities": [0.003365387674421072, 0.014391938224434853, 0.6826915740966797, 0.0920320525765419, 0.0960727334022522, 0.1114463359117508], "decide_seconds": 0.08988873099951888, "probabilities_f64_minus_runtime_max_abs": 2.494537698627397e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.003365387612049128, "einstein": 0.014391937382428326, "hypatia": 0.6826915491513027, "ibn_sina": 0.09203205090481084, "sagan": 0.09607274316021788, "socrates": 0.11144633178919115}, "confidence": 0.42517543281735126, "seconds": 0.0940550980012631, "prompt_tokens": 323, "eval_tokens": 1, "server_latency_ms": 94.03084800214856, "prompt_eval_ms": null, "total_ms": 94.0550980012631, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:01Z"}
{"mode": "kev-choice", "prompt_sha256": "ee3137cb7f1a66a6f4433d42bfd7c7a56e7450e35fcb2e2ab1ca53452e4d5595", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.015756024690490943, "einstein": 0.11551614597239827, "hypatia": 0.3103471663005071, "ibn_sina": 0.21246670891472233, "sagan": 0.16418154744328725, "socrates": 0.18173240667859403}, "confidence": 0.09949734228831242, "choice": "hypatia"}}, "usage": {"input_tokens": 342, "output_tokens": 1}, "latency_ms": 94.85771199979354, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.625, -1.6328125, -0.64453125, -1.0234375, -1.28125, -1.1796875], "runtime_probabilities": [0.015756024047732353, 0.11551614850759506, 0.31034716963768005, 0.21246670186519623, 0.16418154537677765, 0.1817324012517929], "decide_seconds": 0.09066196500134538, "probabilities_f64_minus_runtime_max_abs": 7.049526101754466e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.015756024690490943, "einstein": 0.11551614597239827, "hypatia": 0.3103471663005071, "ibn_sina": 0.21246670891472233, "sagan": 0.16418154744328725, "socrates": 0.18173240667859403}, "confidence": 0.09949734228831242, "seconds": 0.09488389200123493, "prompt_tokens": 342, "eval_tokens": 1, "server_latency_ms": 94.85771199979354, "prompt_eval_ms": null, "total_ms": 94.88389200123493, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:01Z"}
{"mode": "kev-choice", "prompt_sha256": "fef9b4a6bb369848c472d86f9cddd6d23a488ef8c4f0ddf47669b6b7d3933f4e", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.01347151747679994, "einstein": 0.04776067681310916, "hypatia": 0.5824126415544721, "ibn_sina": 0.14258704283079757, "sagan": 0.11728890212139144, "socrates": 0.09647921920342974}, "confidence": 0.28961788303277824, "choice": "hypatia"}}, "usage": {"input_tokens": 339, "output_tokens": 1}, "latency_ms": 95.09757100022398, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.546875, -2.28125, 0.2197265625, -1.1875, -1.3828125, -1.578125], "runtime_probabilities": [0.013471517711877823, 0.047760676592588425, 0.5824126601219177, 0.14258703589439392, 0.11728889495134354, 0.09647922217845917], "decide_seconds": 0.09075446399947396, "probabilities_f64_minus_runtime_max_abs": 1.8567445581041397e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.01347151747679994, "einstein": 0.04776067681310916, "hypatia": 0.5824126415544721, "ibn_sina": 0.14258704283079757, "sagan": 0.11728890212139144, "socrates": 0.09647921920342974}, "confidence": 0.28961788303277824, "seconds": 0.09512444099891582, "prompt_tokens": 339, "eval_tokens": 1, "server_latency_ms": 95.09757100022398, "prompt_eval_ms": null, "total_ms": 95.12444099891582, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:01Z"}
{"mode": "kev-choice", "prompt_sha256": "a1a0666b3c7760ecf67320bc79e8c40eabd3038760e7b0d833c83d02d9857e7c", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.022144842345027782, "einstein": 0.0917878208308319, "hypatia": 0.36730713131302567, "ibn_sina": 0.21010441430742988, "sagan": 0.15251944835864542, "socrates": 0.15613634284503933}, "confidence": 0.12040663547948793, "choice": "hypatia"}}, "usage": {"input_tokens": 351, "output_tokens": 1}, "latency_ms": 94.70858399799909, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.375, -1.953125, -0.56640625, -1.125, -1.4453125, -1.421875], "runtime_probabilities": [0.022144842892885208, 0.09178782254457474, 0.3673071265220642, 0.21010439097881317, 0.15251944959163666, 0.15613633394241333], "decide_seconds": 0.09040752600049018, "probabilities_f64_minus_runtime_max_abs": 2.332861670506503e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.022144842345027782, "einstein": 0.0917878208308319, "hypatia": 0.36730713131302567, "ibn_sina": 0.21010441430742988, "sagan": 0.15251944835864542, "socrates": 0.15613634284503933}, "confidence": 0.12040663547948793, "seconds": 0.09474062400113326, "prompt_tokens": 351, "eval_tokens": 1, "server_latency_ms": 94.70858399799909, "prompt_eval_ms": null, "total_ms": 94.74062400113326, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:01Z"}
{"mode": "kev-choice", "prompt_sha256": "5d38405aad00196bf315b3f02a4541a96ee3b0b544e6d7b01ee62e4201af80de", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.09172853707733877, "einstein": 0.056955543860476625, "hypatia": 0.2903761876844471, "ibn_sina": 0.22351053094300236, "sagan": 0.1687146002173676, "socrates": 0.1687146002173676}, "confidence": 0.06417882409421849, "choice": "hypatia"}}, "usage": {"input_tokens": 393, "output_tokens": 1}, "latency_ms": 97.5426620025246, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-1.8671875, -2.34375, -0.71484375, -0.9765625, -1.2578125, -1.2578125], "runtime_probabilities": [0.09172853082418442, 0.05695554241538048, 0.2903761863708496, 0.2235105335712433, 0.16871459782123566, 0.16871459782123566], "decide_seconds": 0.09288962799837464, "probabilities_f64_minus_runtime_max_abs": 6.253154355428947e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.09172853707733877, "einstein": 0.056955543860476625, "hypatia": 0.2903761876844471, "ibn_sina": 0.22351053094300236, "sagan": 0.1687146002173676, "socrates": 0.1687146002173676}, "confidence": 0.06417882409421849, "seconds": 0.09756915199977811, "prompt_tokens": 393, "eval_tokens": 1, "server_latency_ms": 97.5426620025246, "prompt_eval_ms": null, "total_ms": 97.56915199977811, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:01Z"}
{"mode": "kev-choice", "prompt_sha256": "1a1a9d479c9f9382fc6b8f35f6aa832428b8ff9246586de0a5a16d82b97ed0d3", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.03018857199506739, "einstein": 0.12910014771029496, "hypatia": 0.2733050148473932, "ibn_sina": 0.2301458968631807, "sagan": 0.15941740354860387, "socrates": 0.17784296503545988}, "confidence": 0.07219113773451202, "choice": "hypatia"}}, "usage": {"input_tokens": 317, "output_tokens": 1}, "latency_ms": 91.32471900011296, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.09375, -1.640625, -0.890625, -1.0625, -1.4296875, -1.3203125], "runtime_probabilities": [0.03018856979906559, 0.12910014390945435, 0.2733049988746643, 0.2301459014415741, 0.15941740572452545, 0.17784294486045837], "decide_seconds": 0.08729735099768732, "probabilities_f64_minus_runtime_max_abs": 2.017500150275353e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.03018857199506739, "einstein": 0.12910014771029496, "hypatia": 0.2733050148473932, "ibn_sina": 0.2301458968631807, "sagan": 0.15941740354860387, "socrates": 0.17784296503545988}, "confidence": 0.07219113773451202, "seconds": 0.09135116900142748, "prompt_tokens": 317, "eval_tokens": 1, "server_latency_ms": 91.32471900011296, "prompt_eval_ms": null, "total_ms": 91.35116900142748, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:01Z"}
{"mode": "kev-choice", "prompt_sha256": "0eaa53cd7e68184de0b50355e7ae62c609c53927e1fff8511bf6410b513fd76c", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0023909570081816094, "einstein": 0.02017625075709258, "hypatia": 0.673385994033283, "ibn_sina": 0.07467160934861757, "sagan": 0.11144047397290766, "socrates": 0.1179347148799176}, "confidence": 0.4140712332408728, "choice": "hypatia"}}, "usage": {"input_tokens": 352, "output_tokens": 1}, "latency_ms": 98.97181099950103, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-4.125, -1.9921875, 1.515625, -0.68359375, -0.283203125, -0.2265625], "runtime_probabilities": [0.0023909572046250105, 0.020176250487565994, 0.6733860373497009, 0.07467160373926163, 0.11144047975540161, 0.11793472617864609], "decide_seconds": 0.09465983400150435, "probabilities_f64_minus_runtime_max_abs": 4.331641789701024e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0023909570081816094, "einstein": 0.02017625075709258, "hypatia": 0.673385994033283, "ibn_sina": 0.07467160934861757, "sagan": 0.11144047397290766, "socrates": 0.1179347148799176}, "confidence": 0.4140712332408728, "seconds": 0.09899855100229615, "prompt_tokens": 352, "eval_tokens": 1, "server_latency_ms": 98.97181099950103, "prompt_eval_ms": null, "total_ms": 98.99855100229615, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:02Z"}
{"mode": "kev-choice", "prompt_sha256": "c461c8ab53ff01b5a96bd7ece4257ee0a3381a919c1f261f77bf828da81d00de", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0038708365764845367, "einstein": 0.7852252767112096, "hypatia": 0.0033630410979826842, "ibn_sina": 0.053330986811564826, "sagan": 0.06210715988292218, "socrates": 0.09210269891983612}, "confidence": 0.5651907588263008, "choice": "einstein"}}, "usage": {"input_tokens": 365, "output_tokens": 1}, "latency_ms": 100.39745000176481, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.09375, 2.21875, -3.234375, -0.470703125, -0.318359375, 0.07568359375], "runtime_probabilities": [0.0038708364591002464, 0.7852252721786499, 0.0033630412071943283, 0.053330983966588974, 0.062107156962156296, 0.09210270643234253], "decide_seconds": 0.09555037699828972, "probabilities_f64_minus_runtime_max_abs": 7.512506408624553e-09}}}, "choice": "einstein", "probs": {"darwin": 0.0038708365764845367, "einstein": 0.7852252767112096, "hypatia": 0.0033630410979826842, "ibn_sina": 0.053330986811564826, "sagan": 0.06210715988292218, "socrates": 0.09210269891983612}, "confidence": 0.5651907588263008, "seconds": 0.10042512899963185, "prompt_tokens": 365, "eval_tokens": 1, "server_latency_ms": 100.39745000176481, "prompt_eval_ms": null, "total_ms": 100.42512899963185, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:02Z"}
{"mode": "kev-choice", "prompt_sha256": "9bf1321b83d4fa4a12fee84c88e428fa1366c4478d4eb1e3b333869cc146f6bf", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.024254809484612894, "einstein": 0.3670183335489972, "hypatia": 0.06909552220303634, "ibn_sina": 0.17101305985052292, "sagan": 0.17101305985052292, "socrates": 0.19760521506230766}, "confidence": 0.1253490420971448, "choice": "einstein"}}, "usage": {"input_tokens": 390, "output_tokens": 1}, "latency_ms": 99.2625479993876, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.09375, -0.376953125, -2.046875, -1.140625, -1.140625, -0.99609375], "runtime_probabilities": [0.02425481006503105, 0.36701834201812744, 0.06909552961587906, 0.17101307213306427, 0.17101307213306427, 0.19760523736476898], "decide_seconds": 0.09421567699973821, "probabilities_f64_minus_runtime_max_abs": 2.2302461322176725e-08}}}, "choice": "einstein", "probs": {"darwin": 0.024254809484612894, "einstein": 0.3670183335489972, "hypatia": 0.06909552220303634, "ibn_sina": 0.17101305985052292, "sagan": 0.17101305985052292, "socrates": 0.19760521506230766}, "confidence": 0.1253490420971448, "seconds": 0.09929475799799548, "prompt_tokens": 390, "eval_tokens": 1, "server_latency_ms": 99.2625479993876, "prompt_eval_ms": null, "total_ms": 99.29475799799548, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:02Z"}
{"mode": "kev-choice", "prompt_sha256": "eff8abe6388ec33a313b640330e601559224476208f89d3386053c3801501032", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.03419059365193157, "einstein": 0.14851716886629587, "hypatia": 0.19752335229525297, "ibn_sina": 0.24012732819901877, "sagan": 0.1450767723550207, "socrates": 0.23456478463248023}, "confidence": 0.06138615146326787, "choice": "ibn_sina"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 96.10967300250195, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.703125, -1.234375, -0.94921875, -0.75390625, -1.2578125, -0.77734375], "runtime_probabilities": [0.0341905914247036, 0.1485171616077423, 0.1975233554840088, 0.2401273250579834, 0.14507676661014557, 0.23456478118896484], "decide_seconds": 0.09139492900067125, "probabilities_f64_minus_runtime_max_abs": 7.258553563760373e-09}}}, "choice": "ibn_sina", "probs": {"darwin": 0.03419059365193157, "einstein": 0.14851716886629587, "hypatia": 0.19752335229525297, "ibn_sina": 0.24012732819901877, "sagan": 0.1450767723550207, "socrates": 0.23456478463248023}, "confidence": 0.06138615146326787, "seconds": 0.0961408430011943, "prompt_tokens": 371, "eval_tokens": 1, "server_latency_ms": 96.10967300250195, "prompt_eval_ms": null, "total_ms": 96.1408430011943, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:02Z"}
{"mode": "kev-choice", "prompt_sha256": "c5ad5e3e45e478abdf23a6ab5e1d56e3cd7dd44483f5f3db007ff88291281086", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.3712316746761281, "einstein": 0.05183555624342561, "hypatia": 0.13031457403946795, "ibn_sina": 0.11590418144738085, "sagan": 0.12729584018320036, "socrates": 0.2034181734103972}, "confidence": 0.09422102394097531, "choice": "darwin"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 99.08459000143921, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-0.09375, -2.0625, -1.140625, -1.2578125, -1.1640625, -0.6953125], "runtime_probabilities": [0.37123167514801025, 0.0518355555832386, 0.13031458854675293, 0.11590418964624405, 0.1272958368062973, 0.20341818034648895], "decide_seconds": 0.0939936190006847, "probabilities_f64_minus_runtime_max_abs": 1.4507284984421887e-08}}}, "choice": "darwin", "probs": {"darwin": 0.3712316746761281, "einstein": 0.05183555624342561, "hypatia": 0.13031457403946795, "ibn_sina": 0.11590418144738085, "sagan": 0.12729584018320036, "socrates": 0.2034181734103972}, "confidence": 0.09422102394097531, "seconds": 0.0991138499994122, "prompt_tokens": 395, "eval_tokens": 1, "server_latency_ms": 99.08459000143921, "prompt_eval_ms": null, "total_ms": 99.1138499994122, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:02Z"}
{"mode": "kev-choice", "prompt_sha256": "43311265711e35aea9915fd2616e6caa06d83a03a87905bc4b8ee72a6624f423", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0037520823360856653, "einstein": 0.008455448068857249, "hypatia": 0.6930214669077073, "ibn_sina": 0.12461645534328779, "sagan": 0.10816297316395629, "socrates": 0.061991574180105743}, "confidence": 0.44863829631252317, "choice": "hypatia"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 91.53690799939795, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.6875, -2.875, 1.53125, -0.1845703125, -0.326171875, -0.8828125], "runtime_probabilities": [0.0037520823534578085, 0.008455447852611542, 0.6930214762687683, 0.12461647391319275, 0.10816297680139542, 0.061991579830646515], "decide_seconds": 0.08776160700290347, "probabilities_f64_minus_runtime_max_abs": 1.8569904960963335e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0037520823360856653, "einstein": 0.008455448068857249, "hypatia": 0.6930214669077073, "ibn_sina": 0.12461645534328779, "sagan": 0.10816297316395629, "socrates": 0.061991574180105743}, "confidence": 0.44863829631252317, "seconds": 0.09156159800113528, "prompt_tokens": 293, "eval_tokens": 1, "server_latency_ms": 91.53690799939795, "prompt_eval_ms": null, "total_ms": 91.56159800113528, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:02Z"}
{"mode": "kev-choice", "prompt_sha256": "131fc24eee14e0fad8c3f1bb63f44e6e9b71d0d8e5ad9dfd0ceb8f983efc446b", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.008112673687041461, "einstein": 0.0325915317254639, "hypatia": 0.49606588055578704, "ibn_sina": 0.14268151685557018, "sagan": 0.17043421161060107, "socrates": 0.1501141855655363}, "confidence": 0.2395939837282075, "choice": "hypatia"}}, "usage": {"input_tokens": 311, "output_tokens": 1}, "latency_ms": 91.54950700030895, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.5, -2.109375, 0.61328125, -0.6328125, -0.455078125, -0.58203125], "runtime_probabilities": [0.00811267364770174, 0.032591529190540314, 0.4960658848285675, 0.1426815241575241, 0.17043420672416687, 0.15011419355869293], "decide_seconds": 0.0876697880012216, "probabilities_f64_minus_runtime_max_abs": 7.993156636576515e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.008112673687041461, "einstein": 0.0325915317254639, "hypatia": 0.49606588055578704, "ibn_sina": 0.14268151685557018, "sagan": 0.17043421161060107, "socrates": 0.1501141855655363}, "confidence": 0.2395939837282075, "seconds": 0.09157440700073494, "prompt_tokens": 311, "eval_tokens": 1, "server_latency_ms": 91.54950700030895, "prompt_eval_ms": null, "total_ms": 91.57440700073494, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:02Z"}
{"mode": "kev-choice", "prompt_sha256": "34f4066062175b70b87d8fd7b2f00337e29a272948be077081e7cd483f164ddf", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.013397383308216092, "einstein": 0.07831071290764996, "hypatia": 0.3309908271256923, "ibn_sina": 0.21454015258360043, "sagan": 0.17925500704263708, "socrates": 0.1835059170322041}, "confidence": 0.12225672926724185, "choice": "hypatia"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 95.6821959989611, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.828125, -2.0625, -0.62109375, -1.0546875, -1.234375, -1.2109375], "runtime_probabilities": [0.01339738443493843, 0.07831072062253952, 0.33099085092544556, 0.21454016864299774, 0.17925500869750977, 0.18350592255592346], "decide_seconds": 0.09113070200328366, "probabilities_f64_minus_runtime_max_abs": 2.3799753279263314e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.013397383308216092, "einstein": 0.07831071290764996, "hypatia": 0.3309908271256923, "ibn_sina": 0.21454015258360043, "sagan": 0.17925500704263708, "socrates": 0.1835059170322041}, "confidence": 0.12225672926724185, "seconds": 0.09570972599976812, "prompt_tokens": 371, "eval_tokens": 1, "server_latency_ms": 95.6821959989611, "prompt_eval_ms": null, "total_ms": 95.70972599976812, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:02Z"}
{"mode": "kev-choice", "prompt_sha256": "278ce52f1fb5806bd53426fc0d36fb0841720bd0af69b13c82171a1959ec2442", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0075329641808127715, "einstein": 0.6728188421649217, "hypatia": 0.017514666468762072, "ibn_sina": 0.0746087180156661, "sagan": 0.09961570808602921, "socrates": 0.12790910108380807}, "confidence": 0.40799454983729033, "choice": "einstein"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 95.18454000135534, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.234375, 1.2578125, -2.390625, -0.94140625, -0.65234375, -0.40234375], "runtime_probabilities": [0.007532964460551739, 0.6728187799453735, 0.0175146646797657, 0.0746087059378624, 0.09961570054292679, 0.1279090940952301], "decide_seconds": 0.0907512339981622, "probabilities_f64_minus_runtime_max_abs": 6.221954818030895e-08}}}, "choice": "einstein", "probs": {"darwin": 0.0075329641808127715, "einstein": 0.6728188421649217, "hypatia": 0.017514666468762072, "ibn_sina": 0.0746087180156661, "sagan": 0.09961570808602921, "socrates": 0.12790910108380807}, "confidence": 0.40799454983729033, "seconds": 0.09521346900146455, "prompt_tokens": 355, "eval_tokens": 1, "server_latency_ms": 95.18454000135534, "prompt_eval_ms": null, "total_ms": 95.21346900146455, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:02Z"}
{"mode": "kev-choice", "prompt_sha256": "1ccbf400ac8a33b8f58b4d0235d5d5224724fdf4a03d6c8009022a3d8edce728", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.10008763562550757, "einstein": 0.05929996160110061, "hypatia": 0.27527343381841557, "ibn_sina": 0.21946711608492572, "sagan": 0.17293592643502526, "socrates": 0.17293592643502526}, "confidence": 0.055238701698083426, "choice": "hypatia"}}, "usage": {"input_tokens": 371, "output_tokens": 1}, "latency_ms": 94.40673599965521, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-1.5546875, -2.078125, -0.54296875, -0.76953125, -1.0078125, -1.0078125], "runtime_probabilities": [0.10008764266967773, 0.05929996445775032, 0.2752734422683716, 0.2194671332836151, 0.17293593287467957, 0.17293593287467957], "decide_seconds": 0.0899316099967109, "probabilities_f64_minus_runtime_max_abs": 1.7198689389319455e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.10008763562550757, "einstein": 0.05929996160110061, "hypatia": 0.27527343381841557, "ibn_sina": 0.21946711608492572, "sagan": 0.17293592643502526, "socrates": 0.17293592643502526}, "confidence": 0.055238701698083426, "seconds": 0.09443375599948922, "prompt_tokens": 371, "eval_tokens": 1, "server_latency_ms": 94.40673599965521, "prompt_eval_ms": null, "total_ms": 94.43375599948922, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:02Z"}
{"mode": "kev-choice", "prompt_sha256": "ead36546ab9734fefac1b80c3330c225213caa33b146a86d9dab333bbe6d478c", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.015808040484074012, "einstein": 0.06655433634944169, "hypatia": 0.34736025908157964, "ibn_sina": 0.24155113794189953, "sagan": 0.15474346820596563, "socrates": 0.1739827579370395}, "confidence": 0.135276672343833, "choice": "hypatia"}}, "usage": {"input_tokens": 346, "output_tokens": 1}, "latency_ms": 94.35191699958523, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.6875, -2.25, -0.59765625, -0.9609375, -1.40625, -1.2890625], "runtime_probabilities": [0.015808040276169777, 0.06655433028936386, 0.347360223531723, 0.24155111610889435, 0.15474346280097961, 0.17398275434970856], "decide_seconds": 0.09012637900013942, "probabilities_f64_minus_runtime_max_abs": 3.5549856614025543e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.015808040484074012, "einstein": 0.06655433634944169, "hypatia": 0.34736025908157964, "ibn_sina": 0.24155113794189953, "sagan": 0.15474346820596563, "socrates": 0.1739827579370395}, "confidence": 0.135276672343833, "seconds": 0.09437798599901726, "prompt_tokens": 346, "eval_tokens": 1, "server_latency_ms": 94.35191699958523, "prompt_eval_ms": null, "total_ms": 94.37798599901726, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:03Z"}
{"mode": "kev-choice", "prompt_sha256": "66aa66c272b747d081a054700b83be9e3bd9d52d659c476b286362026ac9636f", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.05352311915008209, "einstein": 0.1477820708327431, "hypatia": 0.249429471084447, "ibn_sina": 0.19425586740165365, "sagan": 0.12251563986534247, "socrates": 0.23249383166573176}, "confidence": 0.051032591509729164, "choice": "hypatia"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 97.74948099948233, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.78125, -1.765625, -1.2421875, -1.4921875, -1.953125, -1.3125], "runtime_probabilities": [0.053523119539022446, 0.1477820724248886, 0.24942946434020996, 0.19425585865974426, 0.12251563370227814, 0.2324938178062439], "decide_seconds": 0.09301329599838937, "probabilities_f64_minus_runtime_max_abs": 1.3859487862921682e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.05352311915008209, "einstein": 0.1477820708327431, "hypatia": 0.249429471084447, "ibn_sina": 0.19425586740165365, "sagan": 0.12251563986534247, "socrates": 0.23249383166573176}, "confidence": 0.051032591509729164, "seconds": 0.09777621000102954, "prompt_tokens": 404, "eval_tokens": 1, "server_latency_ms": 97.74948099948233, "prompt_eval_ms": null, "total_ms": 97.77621000102954, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:03Z"}
{"mode": "kev-choice", "prompt_sha256": "985b8f250ac5989de5ec078b274393d5d7d4a02955eaad8be4423d5426ec6549", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.02853841081344147, "einstein": 0.4455447998643686, "hypatia": 0.04930995204514619, "ibn_sina": 0.14047098116880685, "sagan": 0.12690495592214096, "socrates": 0.20923090018609597}, "confidence": 0.17673593888241002, "choice": "einstein"}}, "usage": {"input_tokens": 309, "output_tokens": 1}, "latency_ms": 94.81915300057153, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.96875, -0.220703125, -2.421875, -1.375, -1.4765625, -0.9765625], "runtime_probabilities": [0.028538411483168602, 0.4455448091030121, 0.049309950321912766, 0.1404709815979004, 0.1269049495458603, 0.20923089981079102], "decide_seconds": 0.09094336199996178, "probabilities_f64_minus_runtime_max_abs": 9.238643461362273e-09}}}, "choice": "einstein", "probs": {"darwin": 0.02853841081344147, "einstein": 0.4455447998643686, "hypatia": 0.04930995204514619, "ibn_sina": 0.14047098116880685, "sagan": 0.12690495592214096, "socrates": 0.20923090018609597}, "confidence": 0.17673593888241002, "seconds": 0.09484412199890357, "prompt_tokens": 309, "eval_tokens": 1, "server_latency_ms": 94.81915300057153, "prompt_eval_ms": null, "total_ms": 94.84412199890357, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:03Z"}
{"mode": "kev-choice", "prompt_sha256": "5ded00af7070527bfe7a3d8e16ee1bf2e306fa18d917c9c2a27b186fa4e3535d", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0036966612050155557, "einstein": 0.02373154166460106, "hypatia": 0.6617779225565286, "ibn_sina": 0.0975990280865754, "sagan": 0.11635534006301394, "socrates": 0.09683950642426543}, "confidence": 0.3937994459778118, "choice": "hypatia"}}, "usage": {"input_tokens": 288, "output_tokens": 1}, "latency_ms": 91.00716200191528, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-4.0625, -2.203125, 1.125, -0.7890625, -0.61328125, -0.796875], "runtime_probabilities": [0.0036966612096875906, 0.023731540888547897, 0.661777913570404, 0.09759902209043503, 0.11635534465312958, 0.09683949500322342], "decide_seconds": 0.08729333000155748, "probabilities_f64_minus_runtime_max_abs": 1.1421042014903016e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0036966612050155557, "einstein": 0.02373154166460106, "hypatia": 0.6617779225565286, "ibn_sina": 0.0975990280865754, "sagan": 0.11635534006301394, "socrates": 0.09683950642426543}, "confidence": 0.3937994459778118, "seconds": 0.09103235200018389, "prompt_tokens": 288, "eval_tokens": 1, "server_latency_ms": 91.00716200191528, "prompt_eval_ms": null, "total_ms": 91.03235200018389, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:03Z"}
{"mode": "kev-choice", "prompt_sha256": "d7c43234b8433372fdf1337f8048ce306f6e3504c3cc8a7f19f1f7827196b61d", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.015813871116474122, "einstein": 0.28583614402630775, "hypatia": 0.1224573092897042, "ibn_sina": 0.2397597723823747, "sagan": 0.16478432104376997, "socrates": 0.17134858214136917}, "confidence": 0.09446569141411021, "choice": "einstein"}}, "usage": {"input_tokens": 327, "output_tokens": 1}, "latency_ms": 94.24154799853568, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.546875, -0.65234375, -1.5, -0.828125, -1.203125, -1.1640625], "runtime_probabilities": [0.01581387035548687, 0.2858361303806305, 0.12245731800794601, 0.23975975811481476, 0.16478432714939117, 0.17134857177734375], "decide_seconds": 0.0901701079965278, "probabilities_f64_minus_runtime_max_abs": 1.4267559939051822e-08}}}, "choice": "einstein", "probs": {"darwin": 0.015813871116474122, "einstein": 0.28583614402630775, "hypatia": 0.1224573092897042, "ibn_sina": 0.2397597723823747, "sagan": 0.16478432104376997, "socrates": 0.17134858214136917}, "confidence": 0.09446569141411021, "seconds": 0.09426842600078089, "prompt_tokens": 327, "eval_tokens": 1, "server_latency_ms": 94.24154799853568, "prompt_eval_ms": null, "total_ms": 94.26842600078089, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:03Z"}
{"mode": "kev-choice", "prompt_sha256": "16b9030c5ff61c8c38b3d8bfb022c702ef1463fba62f7e7ee16a0d11f701bcac", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.1000130196150541, "einstein": 0.1443856541690808, "hypatia": 0.11421845352045641, "ibn_sina": 0.18252056902608724, "sagan": 0.20045962876188042, "socrates": 0.25840267490744095}, "confidence": 0.028994454920387547, "choice": "socrates"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 98.37347600114299, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-1.890625, -1.5234375, -1.7578125, -1.2890625, -1.1953125, -0.94140625], "runtime_probabilities": [0.10001301765441895, 0.14438563585281372, 0.11421845108270645, 0.1825205534696579, 0.20045964419841766, 0.25840267539024353], "decide_seconds": 0.09328930400079116, "probabilities_f64_minus_runtime_max_abs": 1.831626708637124e-08}}}, "choice": "socrates", "probs": {"darwin": 0.1000130196150541, "einstein": 0.1443856541690808, "hypatia": 0.11421845352045641, "ibn_sina": 0.18252056902608724, "sagan": 0.20045962876188042, "socrates": 0.25840267490744095}, "confidence": 0.028994454920387547, "seconds": 0.09840146599890431, "prompt_tokens": 409, "eval_tokens": 1, "server_latency_ms": 98.37347600114299, "prompt_eval_ms": null, "total_ms": 98.40146599890431, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:03Z"}
{"mode": "kev-choice", "prompt_sha256": "6986af98fe6017d32e0ca28fc0acdaa56fd3a5c9618bb0433d6664286443c39a", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.007355124679488659, "einstein": 0.03145387872949974, "hypatia": 0.5729887342769416, "ibn_sina": 0.1303727044101661, "sagan": 0.14544126171849558, "socrates": 0.11238829618540826}, "confidence": 0.2991746120158899, "choice": "hypatia"}}, "usage": {"input_tokens": 331, "output_tokens": 1}, "latency_ms": 94.28574699995806, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-4.15625, -2.703125, 0.19921875, -1.28125, -1.171875, -1.4296875], "runtime_probabilities": [0.007355124689638615, 0.031453877687454224, 0.572988748550415, 0.13037270307540894, 0.14544124901294708, 0.11238829791545868], "decide_seconds": 0.09017026799847372, "probabilities_f64_minus_runtime_max_abs": 1.427347340321461e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.007355124679488659, "einstein": 0.03145387872949974, "hypatia": 0.5729887342769416, "ibn_sina": 0.1303727044101661, "sagan": 0.14544126171849558, "socrates": 0.11238829618540826}, "confidence": 0.2991746120158899, "seconds": 0.09431121699890355, "prompt_tokens": 331, "eval_tokens": 1, "server_latency_ms": 94.28574699995806, "prompt_eval_ms": null, "total_ms": 94.31121699890355, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:03Z"}
{"mode": "kev-choice", "prompt_sha256": "2b2181d544052beef5cbba60e7b4b98d112d0681cb2435766207b405755d7541", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.048821185977083105, "einstein": 0.036280824736340214, "hypatia": 0.4242291394651427, "ibn_sina": 0.2329128519238818, "sagan": 0.12083359577106623, "socrates": 0.1369224021264859}, "confidence": 0.16367179135305643, "choice": "hypatia"}}, "usage": {"input_tokens": 360, "output_tokens": 1}, "latency_ms": 94.40504500162206, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.625, -2.921875, -0.462890625, -1.0625, -1.71875, -1.59375], "runtime_probabilities": [0.048821184784173965, 0.03628082200884819, 0.42422911524772644, 0.2329128384590149, 0.12083358317613602, 0.13692240417003632], "decide_seconds": 0.09007494899924495, "probabilities_f64_minus_runtime_max_abs": 2.4217416239036282e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.048821185977083105, "einstein": 0.036280824736340214, "hypatia": 0.4242291394651427, "ibn_sina": 0.2329128519238818, "sagan": 0.12083359577106623, "socrates": 0.1369224021264859}, "confidence": 0.16367179135305643, "seconds": 0.09443202500187908, "prompt_tokens": 360, "eval_tokens": 1, "server_latency_ms": 94.40504500162206, "prompt_eval_ms": null, "total_ms": 94.43202500187908, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:03Z"}
{"mode": "kev-choice", "prompt_sha256": "fdf1072fd4e4239b84c8e15f11a2561f2bee8da5077fc190be78e7927dc93520", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.017390962189433517, "einstein": 0.35759038554487216, "hypatia": 0.10007807755362857, "ibn_sina": 0.21352697192219805, "sagan": 0.15023498829019782, "socrates": 0.16117861449966991}, "confidence": 0.11974136564884186, "choice": "einstein"}}, "usage": {"input_tokens": 397, "output_tokens": 1}, "latency_ms": 98.00505800012616, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-4.0625, -1.0390625, -2.3125, -1.5546875, -1.90625, -1.8359375], "runtime_probabilities": [0.017390962690114975, 0.3575904071331024, 0.10007807612419128, 0.21352697908878326, 0.15023499727249146, 0.1611786186695099], "decide_seconds": 0.09330581399990479, "probabilities_f64_minus_runtime_max_abs": 2.1588230258995367e-08}}}, "choice": "einstein", "probs": {"darwin": 0.017390962189433517, "einstein": 0.35759038554487216, "hypatia": 0.10007807755362857, "ibn_sina": 0.21352697192219805, "sagan": 0.15023498829019782, "socrates": 0.16117861449966991}, "confidence": 0.11974136564884186, "seconds": 0.09803242799898726, "prompt_tokens": 397, "eval_tokens": 1, "server_latency_ms": 98.00505800012616, "prompt_eval_ms": null, "total_ms": 98.03242799898726, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:03Z"}
{"mode": "kev-choice", "prompt_sha256": "140f7548b8252bf789cc148366539da6583381dee3acb3f73e59965718c98631", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0032233023870861903, "einstein": 0.9221025288707971, "hypatia": 0.0009234915956310948, "ibn_sina": 0.01568080997338288, "sagan": 0.022904741008104808, "socrates": 0.03516512616499797}, "confidence": 0.7939979025823521, "choice": "einstein"}}, "usage": {"input_tokens": 333, "output_tokens": 1}, "latency_ms": 95.02620099738124, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.46875, 3.1875, -3.71875, -0.88671875, -0.5078125, -0.0791015625], "runtime_probabilities": [0.003223302774131298, 0.9221025705337524, 0.000923491723369807, 0.015680810436606407, 0.022904742509126663, 0.03516513109207153], "decide_seconds": 0.09073884399913368, "probabilities_f64_minus_runtime_max_abs": 4.166295530794173e-08}}}, "choice": "einstein", "probs": {"darwin": 0.0032233023870861903, "einstein": 0.9221025288707971, "hypatia": 0.0009234915956310948, "ibn_sina": 0.01568080997338288, "sagan": 0.022904741008104808, "socrates": 0.03516512616499797}, "confidence": 0.7939979025823521, "seconds": 0.09505326100043021, "prompt_tokens": 333, "eval_tokens": 1, "server_latency_ms": 95.02620099738124, "prompt_eval_ms": null, "total_ms": 95.05326100043021, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:03Z"}
{"mode": "kev-choice", "prompt_sha256": "365b36330a3df72c922382a9becff934f9afb952987aa40279f69f09eb7898a3", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.004877724309898841, "einstein": 0.01784197417674043, "hypatia": 0.6643049334352787, "ibn_sina": 0.0787224819560865, "sagan": 0.12192776523726562, "socrates": 0.11232512088472979}, "confidence": 0.40183516039098055, "choice": "hypatia"}}, "usage": {"input_tokens": 366, "output_tokens": 1}, "latency_ms": 94.66045400040457, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.828125, -2.53125, 1.0859375, -1.046875, -0.609375, -0.69140625], "runtime_probabilities": [0.004877724219113588, 0.017841974273324013, 0.6643049120903015, 0.07872247695922852, 0.12192776054143906, 0.11232512444257736], "decide_seconds": 0.09022652799831121, "probabilities_f64_minus_runtime_max_abs": 2.1344977230164375e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.004877724309898841, "einstein": 0.01784197417674043, "hypatia": 0.6643049334352787, "ibn_sina": 0.0787224819560865, "sagan": 0.12192776523726562, "socrates": 0.11232512088472979}, "confidence": 0.40183516039098055, "seconds": 0.09469309300038731, "prompt_tokens": 366, "eval_tokens": 1, "server_latency_ms": 94.66045400040457, "prompt_eval_ms": null, "total_ms": 94.69309300038731, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:03Z"}
{"mode": "kev-choice", "prompt_sha256": "dafefa8a14e06a9ddc266dbf3d063c14fef94ec25b3599ce335c518cc3b141fc", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.01841249220504391, "einstein": 0.3789648364403494, "hypatia": 0.09572379014645292, "ibn_sina": 0.18236279981136877, "sagan": 0.13711799855457632, "socrates": 0.18741808284220868}, "confidence": 0.12797631122465802, "choice": "einstein"}}, "usage": {"input_tokens": 316, "output_tokens": 1}, "latency_ms": 91.61927699824446, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.21875, -0.1943359375, -1.5703125, -0.92578125, -1.2109375, -0.8984375], "runtime_probabilities": [0.01841249316930771, 0.3789648413658142, 0.09572379291057587, 0.18236280977725983, 0.13711799681186676, 0.18741808831691742], "decide_seconds": 0.08766406799986726, "probabilities_f64_minus_runtime_max_abs": 9.965891056840093e-09}}}, "choice": "einstein", "probs": {"darwin": 0.01841249220504391, "einstein": 0.3789648364403494, "hypatia": 0.09572379014645292, "ibn_sina": 0.18236279981136877, "sagan": 0.13711799855457632, "socrates": 0.18741808284220868}, "confidence": 0.12797631122465802, "seconds": 0.09164399700239301, "prompt_tokens": 316, "eval_tokens": 1, "server_latency_ms": 91.61927699824446, "prompt_eval_ms": null, "total_ms": 91.64399700239301, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:04Z"}
{"mode": "kev-choice", "prompt_sha256": "8f499288921bc7087ba6d4545f39cd67a4d51e8c55324c97852bc84f7e6cdcad", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.03538444095682229, "einstein": 0.034835856047419796, "hypatia": 0.4302296248908501, "ibn_sina": 0.2534134139931363, "sagan": 0.11970402078626943, "socrates": 0.126432643325502}, "confidence": 0.18432544186519273, "choice": "hypatia"}}, "usage": {"input_tokens": 399, "output_tokens": 1}, "latency_ms": 98.0140779975045, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.765625, -2.78125, -0.267578125, -0.796875, -1.546875, -1.4921875], "runtime_probabilities": [0.035384438931941986, 0.03483585640788078, 0.4302296042442322, 0.2534134089946747, 0.11970401555299759, 0.1264326423406601], "decide_seconds": 0.09310635599831585, "probabilities_f64_minus_runtime_max_abs": 2.0646617915343057e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.03538444095682229, "einstein": 0.034835856047419796, "hypatia": 0.4302296248908501, "ibn_sina": 0.2534134139931363, "sagan": 0.11970402078626943, "socrates": 0.126432643325502}, "confidence": 0.18432544186519273, "seconds": 0.09804155800156877, "prompt_tokens": 399, "eval_tokens": 1, "server_latency_ms": 98.0140779975045, "prompt_eval_ms": null, "total_ms": 98.04155800156877, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:04Z"}
{"mode": "kev-choice", "prompt_sha256": "6fff42902f8b1b0e2028d0c6ecb1f95991fbdaf15e78d7d0fbaeb93b8ad7e726", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0040661538301433936, "einstein": 0.8712117405082737, "hypatia": 0.0022455434154364387, "ibn_sina": 0.02757087736694065, "sagan": 0.04763820416258505, "socrates": 0.047267480716620694}, "confidence": 0.6961231659159122, "choice": "einstein"}}, "usage": {"input_tokens": 400, "output_tokens": 1}, "latency_ms": 97.96133800045936, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.71875, 1.6484375, -4.3125, -1.8046875, -1.2578125, -1.265625], "runtime_probabilities": [0.004066153429448605, 0.8712116479873657, 0.0022455432917922735, 0.027570873498916626, 0.04763820394873619, 0.047267477959394455], "decide_seconds": 0.0931396460000542, "probabilities_f64_minus_runtime_max_abs": 9.252090793498979e-08}}}, "choice": "einstein", "probs": {"darwin": 0.0040661538301433936, "einstein": 0.8712117405082737, "hypatia": 0.0022455434154364387, "ibn_sina": 0.02757087736694065, "sagan": 0.04763820416258505, "socrates": 0.047267480716620694}, "confidence": 0.6961231659159122, "seconds": 0.09799050800211262, "prompt_tokens": 400, "eval_tokens": 1, "server_latency_ms": 97.96133800045936, "prompt_eval_ms": null, "total_ms": 97.99050800211262, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:04Z"}
{"mode": "kev-choice", "prompt_sha256": "895034ce1c9e079d133fa61f105a7d7186e0ba55df1f35ec288f4512b472d655", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.017117967140029836, "einstein": 0.3652841328236018, "hypatia": 0.07494024943648146, "ibn_sina": 0.17560789175398062, "sagan": 0.26777005857711617, "socrates": 0.0992797002687901}, "confidence": 0.15207167716004988, "choice": "einstein"}}, "usage": {"input_tokens": 284, "output_tokens": 1}, "latency_ms": 90.75471400137758, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.328125, -0.267578125, -1.8515625, -1.0, -0.578125, -1.5703125], "runtime_probabilities": [0.017117967829108238, 0.36528411507606506, 0.07494024932384491, 0.17560790479183197, 0.26777002215385437, 0.09927970170974731], "decide_seconds": 0.08716544100025203, "probabilities_f64_minus_runtime_max_abs": 3.642326179553734e-08}}}, "choice": "einstein", "probs": {"darwin": 0.017117967140029836, "einstein": 0.3652841328236018, "hypatia": 0.07494024943648146, "ibn_sina": 0.17560789175398062, "sagan": 0.26777005857711617, "socrates": 0.0992797002687901}, "confidence": 0.15207167716004988, "seconds": 0.09077969400095753, "prompt_tokens": 284, "eval_tokens": 1, "server_latency_ms": 90.75471400137758, "prompt_eval_ms": null, "total_ms": 90.77969400095753, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:04Z"}
{"mode": "kev-choice", "prompt_sha256": "30c7304be2a27ed3b3947ee0a061c75bc066da90f34275c77e6af9df4ea31263", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.05693116454031892, "einstein": 0.019369853195089232, "hypatia": 0.5481184076198007, "ibn_sina": 0.10471243706013485, "sagan": 0.12434909574964195, "socrates": 0.14651904183501438}, "confidence": 0.24876252802638577, "choice": "hypatia"}}, "usage": {"input_tokens": 404, "output_tokens": 1}, "latency_ms": 98.033537997253, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.078125, -3.15625, 0.1865234375, -1.46875, -1.296875, -1.1328125], "runtime_probabilities": [0.05693116411566734, 0.019369853660464287, 0.5481184124946594, 0.10471243411302567, 0.12434910237789154, 0.14651905000209808], "decide_seconds": 0.0931291950000741, "probabilities_f64_minus_runtime_max_abs": 8.167083703769507e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.05693116454031892, "einstein": 0.019369853195089232, "hypatia": 0.5481184076198007, "ibn_sina": 0.10471243706013485, "sagan": 0.12434909574964195, "socrates": 0.14651904183501438}, "confidence": 0.24876252802638577, "seconds": 0.09806162799941376, "prompt_tokens": 404, "eval_tokens": 1, "server_latency_ms": 98.033537997253, "prompt_eval_ms": null, "total_ms": 98.06162799941376, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:04Z"}
{"mode": "kev-choice", "prompt_sha256": "dc947051091fa4aa147b9543cf9fbe8970fe11bf085718543ba9583ec117426a", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.007015141314257829, "einstein": 0.07781396065367349, "hypatia": 0.2901038106504958, "ibn_sina": 0.22781630125019392, "sagan": 0.13656775402175508, "socrates": 0.2606830321096239}, "confidence": 0.13389258011252902, "choice": "hypatia"}}, "usage": {"input_tokens": 293, "output_tokens": 1}, "latency_ms": 90.83685299992794, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.796875, -1.390625, -0.07470703125, -0.31640625, -0.828125, -0.181640625], "runtime_probabilities": [0.0070151411928236485, 0.07781396061182022, 0.2901037931442261, 0.22781631350517273, 0.13656775653362274, 0.2606830298900604], "decide_seconds": 0.08712811199802672, "probabilities_f64_minus_runtime_max_abs": 1.7506269711642375e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.007015141314257829, "einstein": 0.07781396065367349, "hypatia": 0.2901038106504958, "ibn_sina": 0.22781630125019392, "sagan": 0.13656775402175508, "socrates": 0.2606830321096239}, "confidence": 0.13389258011252902, "seconds": 0.09086459199897945, "prompt_tokens": 293, "eval_tokens": 1, "server_latency_ms": 90.83685299992794, "prompt_eval_ms": null, "total_ms": 90.86459199897945, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:04Z"}
{"mode": "kev-choice", "prompt_sha256": "ad9c564a4063a18ea249262eb0f3a4cc43a639316ae92fe79784687d6575409f", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.003362053980432611, "einstein": 0.015067680577991417, "hypatia": 0.6767078143514583, "ibn_sina": 0.08807397540112442, "sagan": 0.09710893027117136, "socrates": 0.11967954541782184}, "confidence": 0.4189343744357028, "choice": "hypatia"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 91.28101999885985, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.6875, -2.1875, 1.6171875, -0.421875, -0.32421875, -0.115234375], "runtime_probabilities": [0.003362054005265236, 0.01506768073886633, 0.6767078042030334, 0.0880739763379097, 0.09710893034934998, 0.1196795403957367], "decide_seconds": 0.08752362800078117, "probabilities_f64_minus_runtime_max_abs": 1.014842487823131e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.003362053980432611, "einstein": 0.015067680577991417, "hypatia": 0.6767078143514583, "ibn_sina": 0.08807397540112442, "sagan": 0.09710893027117136, "socrates": 0.11967954541782184}, "confidence": 0.4189343744357028, "seconds": 0.09130562000063946, "prompt_tokens": 302, "eval_tokens": 1, "server_latency_ms": 91.28101999885985, "prompt_eval_ms": null, "total_ms": 91.30562000063946, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:04Z"}
{"mode": "kev-choice", "prompt_sha256": "d5a506422f5cb8f1f1acbdb898efea3980d97e1926e0008e68ee1e8e095eaf1f", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.010142403690042535, "einstein": 0.3046211186226675, "hypatia": 0.1433318427248148, "ibn_sina": 0.15497867400226384, "sagan": 0.20854660060601185, "socrates": 0.17837936035419952}, "confidence": 0.10118217644752714, "choice": "einstein"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 97.77395999844884, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.90625, -0.50390625, -1.2578125, -1.1796875, -0.8828125, -1.0390625], "runtime_probabilities": [0.010142404586076736, 0.3046211302280426, 0.1433318555355072, 0.1549786776304245, 0.20854660868644714, 0.1783793717622757], "decide_seconds": 0.09343101299964474, "probabilities_f64_minus_runtime_max_abs": 1.2810692395781587e-08}}}, "choice": "einstein", "probs": {"darwin": 0.010142403690042535, "einstein": 0.3046211186226675, "hypatia": 0.1433318427248148, "ibn_sina": 0.15497867400226384, "sagan": 0.20854660060601185, "socrates": 0.17837936035419952}, "confidence": 0.10118217644752714, "seconds": 0.09780192999824067, "prompt_tokens": 355, "eval_tokens": 1, "server_latency_ms": 97.77395999844884, "prompt_eval_ms": null, "total_ms": 97.80192999824067, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:04Z"}
{"mode": "kev-choice", "prompt_sha256": "b5054215dae25e7dd6ed0d26aa4f538ffbc9f90176e00fecd90f25354f4f77ee", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.017163469516958248, "einstein": 0.026171165726316396, "hypatia": 0.6038526039302801, "ibn_sina": 0.12101424427218303, "sagan": 0.10596361282309916, "socrates": 0.1258349037311631}, "confidence": 0.31689715022398424, "choice": "hypatia"}}, "usage": {"input_tokens": 409, "output_tokens": 1}, "latency_ms": 98.1798870016064, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.234375, -2.8125, 0.326171875, -1.28125, -1.4140625, -1.2421875], "runtime_probabilities": [0.017163468524813652, 0.026171166449785233, 0.6038525700569153, 0.12101424485445023, 0.10596361756324768, 0.12583491206169128], "decide_seconds": 0.09325330499996198, "probabilities_f64_minus_runtime_max_abs": 3.3873364779779536e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.017163469516958248, "einstein": 0.026171165726316396, "hypatia": 0.6038526039302801, "ibn_sina": 0.12101424427218303, "sagan": 0.10596361282309916, "socrates": 0.1258349037311631}, "confidence": 0.31689715022398424, "seconds": 0.09820759600188467, "prompt_tokens": 409, "eval_tokens": 1, "server_latency_ms": 98.1798870016064, "prompt_eval_ms": null, "total_ms": 98.20759600188467, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:04Z"}
{"mode": "kev-choice", "prompt_sha256": "42e356c1a5a738bdceafa36690a1ba81cb1f60c2454be3eeb7763dc8bc966512", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.027977750881182577, "einstein": 0.029782160786003115, "hypatia": 0.44367018415463827, "ibn_sina": 0.182437852889848, "sagan": 0.12936780939479584, "socrates": 0.18676424189353227}, "confidence": 0.188730488951664, "choice": "hypatia"}}, "usage": {"input_tokens": 291, "output_tokens": 1}, "latency_ms": 91.82603500084952, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.125, -3.0625, -0.361328125, -1.25, -1.59375, -1.2265625], "runtime_probabilities": [0.027977753430604935, 0.029782161116600037, 0.4436701834201813, 0.18243785202503204, 0.12936781346797943, 0.18676424026489258], "decide_seconds": 0.08818878300007782, "probabilities_f64_minus_runtime_max_abs": 4.073183595965091e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.027977750881182577, "einstein": 0.029782160786003115, "hypatia": 0.44367018415463827, "ibn_sina": 0.182437852889848, "sagan": 0.12936780939479584, "socrates": 0.18676424189353227}, "confidence": 0.188730488951664, "seconds": 0.09185079500093707, "prompt_tokens": 291, "eval_tokens": 1, "server_latency_ms": 91.82603500084952, "prompt_eval_ms": null, "total_ms": 91.85079500093707, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:04Z"}
{"mode": "kev-choice", "prompt_sha256": "67dcb6aab5956e3345ea02d36dd1493202dd845c88aa2d1ddf5c08d569e9e261", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.09053967582002816, "einstein": 0.06370261805415582, "hypatia": 0.35253967231175715, "ibn_sina": 0.17181411428212442, "sagan": 0.10421052808347872, "socrates": 0.2171933914484558}, "confidence": 0.09007780334517912, "choice": "hypatia"}}, "usage": {"input_tokens": 361, "output_tokens": 1}, "latency_ms": 94.81983300065622, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-1.9921875, -2.34375, -0.6328125, -1.3515625, -1.8515625, -1.1171875], "runtime_probabilities": [0.09053967893123627, 0.06370261311531067, 0.35253965854644775, 0.17181411385536194, 0.10421052575111389, 0.2171933948993683], "decide_seconds": 0.09041403599985642, "probabilities_f64_minus_runtime_max_abs": 1.3765309392343994e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.09053967582002816, "einstein": 0.06370261805415582, "hypatia": 0.35253967231175715, "ibn_sina": 0.17181411428212442, "sagan": 0.10421052808347872, "socrates": 0.2171933914484558}, "confidence": 0.09007780334517912, "seconds": 0.09484705299837515, "prompt_tokens": 361, "eval_tokens": 1, "server_latency_ms": 94.81983300065622, "prompt_eval_ms": null, "total_ms": 94.84705299837515, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:05Z"}
{"mode": "kev-choice", "prompt_sha256": "14b0361f2a076d27818b3a17225e24173e6c36324e06c05946a2103c8059563e", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.03444255151636518, "einstein": 0.04492155573169114, "hypatia": 0.4876916725644978, "ibn_sina": 0.18619467434901366, "sagan": 0.1094582023466764, "socrates": 0.13729134349175573}, "confidence": 0.20003545649066456, "choice": "hypatia"}}, "usage": {"input_tokens": 343, "output_tokens": 1}, "latency_ms": 94.07275800185744, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.03125, -2.765625, -0.380859375, -1.34375, -1.875, -1.6484375], "runtime_probabilities": [0.03444255143404007, 0.04492156207561493, 0.4876916706562042, 0.18619467318058014, 0.10945820808410645, 0.1372913420200348], "decide_seconds": 0.08990632100176299, "probabilities_f64_minus_runtime_max_abs": 6.343923789065009e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.03444255151636518, "einstein": 0.04492155573169114, "hypatia": 0.4876916725644978, "ibn_sina": 0.18619467434901366, "sagan": 0.1094582023466764, "socrates": 0.13729134349175573}, "confidence": 0.20003545649066456, "seconds": 0.09409841800152208, "prompt_tokens": 343, "eval_tokens": 1, "server_latency_ms": 94.07275800185744, "prompt_eval_ms": null, "total_ms": 94.09841800152208, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:05Z"}
{"mode": "kev-choice", "prompt_sha256": "d60dfdc0fb4ca19c34bc1382c9dd3b8883127f3d35673599556ccf58a577cec6", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0245803539701563, "einstein": 0.0245803539701563, "hypatia": 0.5029557490996576, "ibn_sina": 0.2010412286672583, "sagan": 0.12679511357335768, "socrates": 0.12004720071941387}, "confidence": 0.23723185733742713, "choice": "hypatia"}}, "usage": {"input_tokens": 374, "output_tokens": 1}, "latency_ms": 94.4465449974814, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.203125, -3.203125, -0.1845703125, -1.1015625, -1.5625, -1.6171875], "runtime_probabilities": [0.024580353870987892, 0.024580353870987892, 0.5029557347297668, 0.20104123651981354, 0.12679509818553925, 0.1200471967458725], "decide_seconds": 0.09003692000260344, "probabilities_f64_minus_runtime_max_abs": 1.538781843613002e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0245803539701563, "einstein": 0.0245803539701563, "hypatia": 0.5029557490996576, "ibn_sina": 0.2010412286672583, "sagan": 0.12679511357335768, "socrates": 0.12004720071941387}, "confidence": 0.23723185733742713, "seconds": 0.09447406500112265, "prompt_tokens": 374, "eval_tokens": 1, "server_latency_ms": 94.4465449974814, "prompt_eval_ms": null, "total_ms": 94.47406500112265, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:05Z"}
{"mode": "kev-choice", "prompt_sha256": "1926d0b3635e02750addb5db2c1806192a7739954e8245ae4a55fa4e482c4191", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.012478514872593213, "einstein": 0.6629268188376891, "hypatia": 0.016275051545806312, "ibn_sina": 0.08010861130836135, "sagan": 0.09969677757847407, "socrates": 0.1285142258570759}, "confidence": 0.3916538431103913, "choice": "einstein"}}, "usage": {"input_tokens": 302, "output_tokens": 1}, "latency_ms": 91.001321998192, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.0625, 0.91015625, -2.796875, -1.203125, -0.984375, -0.73046875], "runtime_probabilities": [0.012478513643145561, 0.6629267930984497, 0.01627505198121071, 0.08010861277580261, 0.09969677776098251, 0.12851421535015106], "decide_seconds": 0.08727481000096304, "probabilities_f64_minus_runtime_max_abs": 2.5739239406519232e-08}}}, "choice": "einstein", "probs": {"darwin": 0.012478514872593213, "einstein": 0.6629268188376891, "hypatia": 0.016275051545806312, "ibn_sina": 0.08010861130836135, "sagan": 0.09969677757847407, "socrates": 0.1285142258570759}, "confidence": 0.3916538431103913, "seconds": 0.0910259819975181, "prompt_tokens": 302, "eval_tokens": 1, "server_latency_ms": 91.001321998192, "prompt_eval_ms": null, "total_ms": 91.0259819975181, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:05Z"}
{"mode": "kev-choice", "prompt_sha256": "d89193cb5dc2ab9536b637ce578af9ac6c54d6ee49f7dab2728098cc245c3b79", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.006846928329769381, "einstein": 0.03165987583731007, "hypatia": 0.6115444439461798, "ibn_sina": 0.09751928298873536, "sagan": 0.12424278969789979, "socrates": 0.12818667920010554}, "confidence": 0.33383291273151017, "choice": "hypatia"}}, "usage": {"input_tokens": 323, "output_tokens": 1}, "latency_ms": 94.02799799863715, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.671875, -2.140625, 0.8203125, -1.015625, -0.7734375, -0.7421875], "runtime_probabilities": [0.006846928969025612, 0.03165987879037857, 0.6115444898605347, 0.09751929342746735, 0.12424279749393463, 0.1281866878271103], "decide_seconds": 0.09008961899962742, "probabilities_f64_minus_runtime_max_abs": 4.591435487366624e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.006846928329769381, "einstein": 0.03165987583731007, "hypatia": 0.6115444439461798, "ibn_sina": 0.09751928298873536, "sagan": 0.12424278969789979, "socrates": 0.12818667920010554}, "confidence": 0.33383291273151017, "seconds": 0.09405435799999395, "prompt_tokens": 323, "eval_tokens": 1, "server_latency_ms": 94.02799799863715, "prompt_eval_ms": null, "total_ms": 94.05435799999395, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:05Z"}
{"mode": "kev-choice", "prompt_sha256": "b1bf9094b8e8971fa3e29b32555834ffa5d0e33970229230c1b2ce4615d93c8d", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.03136569051547874, "einstein": 0.04934515756676919, "hypatia": 0.45732858494236256, "ibn_sina": 0.15318588778608827, "sagan": 0.1543873395946506, "socrates": 0.1543873395946506}, "confidence": 0.17448152961022978, "choice": "hypatia"}}, "usage": {"input_tokens": 395, "output_tokens": 1}, "latency_ms": 97.1890139990137, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-2.96875, -2.515625, -0.2890625, -1.3828125, -1.375, -1.375], "runtime_probabilities": [0.031365688890218735, 0.04934515804052353, 0.45732858777046204, 0.1531858891248703, 0.15438733994960785, 0.15438733994960785], "decide_seconds": 0.09262271999978111, "probabilities_f64_minus_runtime_max_abs": 2.8280994768259404e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.03136569051547874, "einstein": 0.04934515756676919, "hypatia": 0.45732858494236256, "ibn_sina": 0.15318588778608827, "sagan": 0.1543873395946506, "socrates": 0.1543873395946506}, "confidence": 0.17448152961022978, "seconds": 0.09721631399952457, "prompt_tokens": 395, "eval_tokens": 1, "server_latency_ms": 97.1890139990137, "prompt_eval_ms": null, "total_ms": 97.21631399952457, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:05Z"}
{"mode": "kev-choice", "prompt_sha256": "7b3f225227115a726e53cd581c75470b8f94abde5ff45d86102ab8830a619d53", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0032181701600712013, "einstein": 0.03053314835520719, "hypatia": 0.5851907864567214, "ibn_sina": 0.10208891760438694, "sagan": 0.14941139622729438, "socrates": 0.12955758119631886}, "confidence": 0.3189280949537636, "choice": "hypatia"}}, "usage": {"input_tokens": 297, "output_tokens": 1}, "latency_ms": 91.08428100080346, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-4.0, -1.75, 1.203125, -0.54296875, -0.162109375, -0.3046875], "runtime_probabilities": [0.003218170255422592, 0.03053314797580242, 0.5851907730102539, 0.10208890587091446, 0.1494114100933075, 0.12955757975578308], "decide_seconds": 0.08737836000000243, "probabilities_f64_minus_runtime_max_abs": 1.386601311548219e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0032181701600712013, "einstein": 0.03053314835520719, "hypatia": 0.5851907864567214, "ibn_sina": 0.10208891760438694, "sagan": 0.14941139622729438, "socrates": 0.12955758119631886}, "confidence": 0.3189280949537636, "seconds": 0.0911086609994527, "prompt_tokens": 297, "eval_tokens": 1, "server_latency_ms": 91.08428100080346, "prompt_eval_ms": null, "total_ms": 91.1086609994527, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:05Z"}
{"mode": "kev-choice", "prompt_sha256": "d88cd71424f272ecebd6dd7042e2a710895408bec112faac2f21b945c91ab755", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.017128993879319175, "einstein": 0.20067969427853655, "hypatia": 0.13160895664493422, "ibn_sina": 0.2576778280667665, "sagan": 0.14119579952176595, "socrates": 0.25170872760867763}, "confidence": 0.08921198663384577, "choice": "ibn_sina"}}, "usage": {"input_tokens": 368, "output_tokens": 1}, "latency_ms": 94.2602569994051, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.671875, -1.2109375, -1.6328125, -0.9609375, -1.5625, -0.984375], "runtime_probabilities": [0.017128994688391685, 0.2006797045469284, 0.1316089630126953, 0.25767782330513, 0.14119580388069153, 0.2517087161540985], "decide_seconds": 0.08985174100234872, "probabilities_f64_minus_runtime_max_abs": 1.1454579118552743e-08}}}, "choice": "ibn_sina", "probs": {"darwin": 0.017128993879319175, "einstein": 0.20067969427853655, "hypatia": 0.13160895664493422, "ibn_sina": 0.2576778280667665, "sagan": 0.14119579952176595, "socrates": 0.25170872760867763}, "confidence": 0.08921198663384577, "seconds": 0.09428585699788528, "prompt_tokens": 368, "eval_tokens": 1, "server_latency_ms": 94.2602569994051, "prompt_eval_ms": null, "total_ms": 94.28585699788528, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:05Z"}
{"mode": "kev-choice", "prompt_sha256": "b3912f0c41f0fc5f8ba5bebb9305b5cb081aa57d59147bf2e054b9e3c92505e2", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.19968188983433977, "einstein": 0.03766596115425775, "hypatia": 0.2574001298039833, "ibn_sina": 0.13146712225079485, "sagan": 0.14214985235652927, "socrates": 0.23163504460009499}, "confidence": 0.06383950330279509, "choice": "hypatia"}}, "usage": {"input_tokens": 298, "output_tokens": 1}, "latency_ms": 90.89577200211352, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-0.70703125, -2.375, -0.453125, -1.125, -1.046875, -0.55859375], "runtime_probabilities": [0.19968189299106598, 0.037665955722332, 0.25740012526512146, 0.13146711885929108, 0.14214983582496643, 0.23163504898548126], "decide_seconds": 0.08711501200014027, "probabilities_f64_minus_runtime_max_abs": 1.6531562840871317e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.19968188983433977, "einstein": 0.03766596115425775, "hypatia": 0.2574001298039833, "ibn_sina": 0.13146712225079485, "sagan": 0.14214985235652927, "socrates": 0.23163504460009499}, "confidence": 0.06383950330279509, "seconds": 0.09092527200118639, "prompt_tokens": 298, "eval_tokens": 1, "server_latency_ms": 90.89577200211352, "prompt_eval_ms": null, "total_ms": 90.92527200118639, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:05Z"}
{"mode": "kev-choice", "prompt_sha256": "087772a504abc6edd03a515893d7186b659ed1ce62bcdf63c6380c519110815b", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.008269407969315965, "einstein": 0.6884467125446215, "hypatia": 0.010291444497134912, "ibn_sina": 0.06978195267865125, "sagan": 0.10192952761097182, "socrates": 0.12128095469930449}, "confidence": 0.43175409327491576, "choice": "einstein"}}, "usage": {"input_tokens": 340, "output_tokens": 1}, "latency_ms": 94.25392699995427, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.0625, 1.359375, -2.84375, -0.9296875, -0.55078125, -0.376953125], "runtime_probabilities": [0.008269408717751503, 0.6884467601776123, 0.010291445069015026, 0.06978195905685425, 0.10192953050136566, 0.1212809607386589], "decide_seconds": 0.0900291500001913, "probabilities_f64_minus_runtime_max_abs": 4.763299077392702e-08}}}, "choice": "einstein", "probs": {"darwin": 0.008269407969315965, "einstein": 0.6884467125446215, "hypatia": 0.010291444497134912, "ibn_sina": 0.06978195267865125, "sagan": 0.10192952761097182, "socrates": 0.12128095469930449}, "confidence": 0.43175409327491576, "seconds": 0.09427958699961891, "prompt_tokens": 340, "eval_tokens": 1, "server_latency_ms": 94.25392699995427, "prompt_eval_ms": null, "total_ms": 94.27958699961891, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:05Z"}
{"mode": "kev-choice", "prompt_sha256": "da39a310f76af967598a18569427bb9a79c40e6c5d35b8fd761b829c3d14d3fb", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.17678472467550757, "einstein": 0.071149016218661, "hypatia": 0.2592376601147893, "ibn_sina": 0.18892290042579113, "sagan": 0.13188944022559965, "socrates": 0.17201625833965128}, "confidence": 0.034950137682772775, "choice": "hypatia"}}, "usage": {"input_tokens": 380, "output_tokens": 1}, "latency_ms": 95.1369499998691, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-0.94140625, -1.8515625, -0.55859375, -0.875, -1.234375, -0.96875], "runtime_probabilities": [0.1767847090959549, 0.07114901393651962, 0.2592376470565796, 0.18892289698123932, 0.1318894326686859, 0.1720162332057953], "decide_seconds": 0.09062000500125578, "probabilities_f64_minus_runtime_max_abs": 2.5133855996450905e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.17678472467550757, "einstein": 0.071149016218661, "hypatia": 0.2592376601147893, "ibn_sina": 0.18892290042579113, "sagan": 0.13188944022559965, "socrates": 0.17201625833965128}, "confidence": 0.034950137682772775, "seconds": 0.09516392000296037, "prompt_tokens": 380, "eval_tokens": 1, "server_latency_ms": 95.1369499998691, "prompt_eval_ms": null, "total_ms": 95.16392000296037, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:05Z"}
{"mode": "kev-choice", "prompt_sha256": "ff6d1510cc4a9f868606a4497f9e4eb74877e054bdde5bac5ba03a41f299221c", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.012094797718782229, "einstein": 0.033920699051984554, "hypatia": 0.5183174640807822, "ibn_sina": 0.15869659465556454, "sagan": 0.1341589672456982, "socrates": 0.14281147724718832}, "confidence": 0.24746947495863536, "choice": "hypatia"}}, "usage": {"input_tokens": 267, "output_tokens": 1}, "latency_ms": 91.51680799914175, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-3.5, -2.46875, 0.2578125, -0.92578125, -1.09375, -1.03125], "runtime_probabilities": [0.012094797566533089, 0.03392069786787033, 0.518317461013794, 0.15869659185409546, 0.13415896892547607, 0.14281146228313446], "decide_seconds": 0.08789291500215768, "probabilities_f64_minus_runtime_max_abs": 1.4964053857147164e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.012094797718782229, "einstein": 0.033920699051984554, "hypatia": 0.5183174640807822, "ibn_sina": 0.15869659465556454, "sagan": 0.1341589672456982, "socrates": 0.14281147724718832}, "confidence": 0.24746947495863536, "seconds": 0.09154147799927159, "prompt_tokens": 267, "eval_tokens": 1, "server_latency_ms": 91.51680799914175, "prompt_eval_ms": null, "total_ms": 91.54147799927159, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:06Z"}
{"mode": "kev-choice", "prompt_sha256": "caa5afacee5ba562467873a4832b48d03710f29e44cf8989baea421877ce095d", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0026452569633492362, "einstein": 0.7626740244254802, "hypatia": 0.009525936260007856, "ibn_sina": 0.051496724885752736, "sagan": 0.05439138140177413, "socrates": 0.11926667606363588}, "confidence": 0.5359977309968272, "choice": "einstein"}}, "usage": {"input_tokens": 315, "output_tokens": 1}, "latency_ms": 91.73217699935776, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-4.5, 1.1640625, -3.21875, -1.53125, -1.4765625, -0.69140625], "runtime_probabilities": [0.00264525692909956, 0.7626740336418152, 0.009525937028229237, 0.0514967255294323, 0.05439138039946556, 0.11926667392253876], "decide_seconds": 0.08775621700260672, "probabilities_f64_minus_runtime_max_abs": 9.216335028483513e-09}}}, "choice": "einstein", "probs": {"darwin": 0.0026452569633492362, "einstein": 0.7626740244254802, "hypatia": 0.009525936260007856, "ibn_sina": 0.051496724885752736, "sagan": 0.05439138140177413, "socrates": 0.11926667606363588}, "confidence": 0.5359977309968272, "seconds": 0.09175790600056644, "prompt_tokens": 315, "eval_tokens": 1, "server_latency_ms": 91.73217699935776, "prompt_eval_ms": null, "total_ms": 91.75790600056644, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:06Z"}
{"mode": "kev-choice", "prompt_sha256": "b85349448634763b9f05d811ce1c6f89af68fc4dc67dea4417ef46bb2e66e79f", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0016289919214302541, "einstein": 0.020315799227073526, "hypatia": 0.7274344343558313, "ibn_sina": 0.058328240188389176, "sagan": 0.09416892807985938, "socrates": 0.09812360622741641}, "confidence": 0.4769701534335772, "choice": "hypatia"}}, "usage": {"input_tokens": 335, "output_tokens": 1}, "latency_ms": 94.23827699720277, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-4.125, -1.6015625, 1.9765625, -0.546875, -0.06787109375, -0.0267333984375], "runtime_probabilities": [0.00162899203132838, 0.020315799862146378, 0.7274344563484192, 0.058328237384557724, 0.09416893124580383, 0.09812361001968384], "decide_seconds": 0.09017791799851693, "probabilities_f64_minus_runtime_max_abs": 2.1992587861952018e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0016289919214302541, "einstein": 0.020315799227073526, "hypatia": 0.7274344343558313, "ibn_sina": 0.058328240188389176, "sagan": 0.09416892807985938, "socrates": 0.09812360622741641}, "confidence": 0.4769701534335772, "seconds": 0.09426438700029394, "prompt_tokens": 335, "eval_tokens": 1, "server_latency_ms": 94.23827699720277, "prompt_eval_ms": null, "total_ms": 94.26438700029394, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:06Z"}
{"mode": "kev-choice", "prompt_sha256": "4de53544ee23e1e93fd3d21ab2a5290ccc1d0e8d4141479fccab952a097e4269", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.11478118454182387, "einstein": 0.08939167640303587, "hypatia": 0.266874249748864, "ibn_sina": 0.19372992941150552, "sagan": 0.12410805193409609, "socrates": 0.21111490796067464}, "confidence": 0.03884929412787619, "choice": "hypatia"}}, "usage": {"input_tokens": 355, "output_tokens": 1}, "latency_ms": 94.39208600088023, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-1.5703125, -1.8203125, -0.7265625, -1.046875, -1.4921875, -0.9609375], "runtime_probabilities": [0.11478118598461151, 0.08939167857170105, 0.2668742537498474, 0.1937299370765686, 0.12410806119441986, 0.21111491322517395], "decide_seconds": 0.09007999899768038, "probabilities_f64_minus_runtime_max_abs": 9.260323771687062e-09}}}, "choice": "hypatia", "probs": {"darwin": 0.11478118454182387, "einstein": 0.08939167640303587, "hypatia": 0.266874249748864, "ibn_sina": 0.19372992941150552, "sagan": 0.12410805193409609, "socrates": 0.21111490796067464}, "confidence": 0.03884929412787619, "seconds": 0.09442018600020674, "prompt_tokens": 355, "eval_tokens": 1, "server_latency_ms": 94.39208600088023, "prompt_eval_ms": null, "total_ms": 94.42018600020674, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:06Z"}
{"mode": "kev-choice", "prompt_sha256": "0f756818bdff4088317d34f8aedc6ab0883679e7961bf2faa0d0a75e163f2296", "n_options": 6, "kev_response": {"model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "answers": {"q": {"type": "choice", "probabilities": {"darwin": 0.0029923141002313565, "einstein": 0.010945430186916688, "hypatia": 0.6986654634611662, "ibn_sina": 0.09566866205828148, "sagan": 0.11272523224968893, "socrates": 0.07900289794371536}, "confidence": 0.44833780050462324, "choice": "hypatia"}}, "usage": {"input_tokens": 328, "output_tokens": 1}, "latency_ms": 94.2220670003735, "apus": {"q": {"effort": "low", "executed_layers": 16, "projection": "candidate_rows", "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": [-4.03125, -2.734375, 1.421875, -0.56640625, -0.40234375, -0.7578125], "runtime_probabilities": [0.002992314053699374, 0.010945430025458336, 0.69866544008255, 0.09566865861415863, 0.11272523552179337, 0.07900290191173553], "decide_seconds": 0.0902234179993684, "probabilities_f64_minus_runtime_max_abs": 2.3378616154090537e-08}}}, "choice": "hypatia", "probs": {"darwin": 0.0029923141002313565, "einstein": 0.010945430186916688, "hypatia": 0.6986654634611662, "ibn_sina": 0.09566866205828148, "sagan": 0.11272523224968893, "socrates": 0.07900289794371536}, "confidence": 0.44833780050462324, "seconds": 0.0942475670017302, "prompt_tokens": 328, "eval_tokens": 1, "server_latency_ms": 94.2220670003735, "prompt_eval_ms": null, "total_ms": 94.2475670017302, "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": "low", "runtime": "openjet_runtime (APUS-OpenJev-v1-4B@422b3741) in-process via deployment/contracts.py (APUS-OpenJev-v1@af276da5), one question per request", "model_revision": "422b3741f8b5c092eeefef847c1ca89d78337d45", "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-4b-low", "model": "apus-ailab/APUS-OpenJev-v1-4B@422b3741", "quant": "bf16 (merged LoRA r=8; full weights incl. the vision tower; tied embeddings)", "card_uuid": "GPU-fp:33a0b4acb8e8", "num_ctx": 8191, "power_cap_w": 300, "seed": 20260921, "stamp": "2026-09-28T16:50:06Z"}