on-policy-self-distillation

gpt-5.6-sol

on-policy-self-distillation — graded by the task's verifier against a measured baseline.

backend: modal
variant: codex azure_ai/gpt-5.6-sol
metric: AIME24 avg@12 accuracy
actions: 636
duration: 8h 10m
recorded: 2026-08-03

run idopsd-aime24-4gpus__3tyLi4e

Everything gpt-5.6-sol did on this task, in order.

download JSON ↗

codexgpt-5.6-sol683 turns47 narration636 reasoning636 action636 observation1 final

#276

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolexec

args

{
  "input": "const r = await tools.write_stdin({session_id:1051,chars:\"\",yield_time_ms:50000,max_output_tokens:250});\ntext(r.output);"
}

exec result

Script running with cell ID 147
Wall time 11.0 seconds
Output:

#277

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolwait

args

{
  "cell_id": "147",
  "yield_time_ms": 50000,
  "max_tokens": 250
}

wait result

Script completed
Wall time 31.9 seconds
Output:

Processed prompts:  29%|▎| 24/84 [08:23<18:25, 18.42s/it, est. speed input: 3.5

#278

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolexec

args

{
  "input": "const r = await tools.write_stdin({session_id:1051,chars:\"\",yield_time_ms:50000,max_output_tokens:250});\ntext(r.output);"
}

exec result

Script running with cell ID 148
Wall time 11.0 seconds
Output:

#279

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolwait

args

{
  "cell_id": "148",
  "yield_time_ms": 50000,
  "max_tokens": 250
}

wait result

Script completed
Wall time 32.7 seconds
Output:

Processed prompts:  25%|▎| 24/96 [09:19<26:20, 21.95s/it, est. speed input: 5.6

#280

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolexec

args

{
  "input": "const r = await tools.write_stdin({session_id:1051,chars:\"\",yield_time_ms:50000,max_output_tokens:250});\ntext(r.output);"
}

exec result

Script running with cell ID 149
Wall time 11.0 seconds
Output:

#281

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolwait

args

{
  "cell_id": "149",
  "yield_time_ms": 50000,
  "max_tokens": 250
}

wait result

Script completed
Wall time 27.6 seconds
Output:

#282

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolexec

args

{
  "input": "const r = await tools.write_stdin({session_id:1051,chars:\"\",yield_time_ms:50000,max_output_tokens:250});\ntext(r.output);"
}

exec result

Script running with cell ID 150
Wall time 11.0 seconds
Output:

#283

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolwait

args

{
  "cell_id": "150",
  "yield_time_ms": 50000,
  "max_tokens": 250
}

wait result

Script completed
Wall time 26.7 seconds
Output:

Processed prompts:  43%|▍| 36/84 [11:24<13:31, 16.90s/it, est. speed input: 4.4

#284

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolexec

args

{
  "input": "const r = await tools.write_stdin({session_id:1051,chars:\"\",yield_time_ms:50000,max_output_tokens:250});\ntext(r.output);"
}

exec result

Script running with cell ID 151
Wall time 11.0 seconds
Output:

#285

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolwait

args

{
  "cell_id": "151",
  "yield_time_ms": 50000,
  "max_tokens": 250
}

wait result

Script completed
Wall time 30.2 seconds
Output:

Processed prompts:  38%|▍| 36/96 [11:39<19:53, 19.90s/it, est. speed input: 7.6
Processed prompts:  50%|▌| 48/96 [11:42<09:43, 12.15s/it, est. speed input: 8.8
Processed prompts:  57%|▌| 48/84 [11:44<06:31, 10.88s/it, est. speed input: 5.7
Processed prompts:  71%|▋| 60/84 [11:53<02:53,  7.22s/it, est. speed input: 8.4
Processed prompts:  86%|▊| 72/84 [12:18<01:05,  5.48s/it, est. speed input: 10.
Processed prompts:  14%|▏| 12/84 [12:42<1:16:13, 63.52s/it, est. speed input: 1

#286

narration

Two v2 shards are moving quickly, while one is again bottlenecked by a near-max-length trace. This is evaluator workload variance, not a training or inference failure; all samples are still being retained for the official metric.

#287

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolexec

args

{
  "input": "const r = await tools.write_stdin({session_id:1051,chars:\"\",yield_time_ms:50000,max_output_tokens:250});\ntext(r.output);"
}

exec result

Script running with cell ID 152
Wall time 11.0 seconds
Output:

#288

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolwait

args

{
  "cell_id": "152",
  "yield_time_ms": 50000,
  "max_tokens": 250
}

wait result

Script completed
Wall time 33.5 seconds
Output:

Processed prompts:  62%|▋| 60/96 [13:03<06:07, 10.21s/it, est. speed input: 9.2
Processed prompts:  75%|▊| 72/96 [13:22<02:54,  7.27s/it, est. speed input: 10.

#289

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolexec

args

{
  "input": "const r = await tools.write_stdin({session_id:1051,chars:\"\",yield_time_ms:50000,max_output_tokens:250});\ntext(r.output);"
}

exec result

Script running with cell ID 153
Wall time 11.0 seconds
Output:

#290

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolwait

args

{
  "cell_id": "153",
  "yield_time_ms": 50000,
  "max_tokens": 250
}

wait result

Script completed
Wall time 33.8 seconds
Output:

Processed prompts:  88%|▉| 84/96 [13:48<01:07,  5.61s/it, est. speed input: 12.
Processed prompts:  29%|▎| 24/84 [13:50<29:28, 29.48s/it, est. speed input: 3.1
Processed prompts: 100%|█| 96/96 [14:06<00:00,  4.30s/it, est. speed input: 17.
Processed prompts: 100%|█| 96/96 [14:06<00:00,  4.30s/it, est. speed input: 17.
Processed prompts: 100%|█| 96/96 [14:06<00:00,  8.82s/it, est. speed input: 17.

#291

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolexec

args

{
  "input": "const r = await tools.write_stdin({session_id:1051,chars:\"\",yield_time_ms:50000,max_output_tokens:250});\ntext(r.output);"
}

exec result

Script running with cell ID 154
Wall time 11.0 seconds
Output:

#292

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolwait

args

{
  "cell_id": "154",
  "yield_time_ms": 50000,
  "max_tokens": 250
}

wait result

Script completed
Wall time 28.3 seconds
Output:

Processed prompts:  43%|▍| 36/84 [15:02<15:01, 18.78s/it, est. speed input: 4.2

#293

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolexec

args

{
  "input": "const r = await tools.write_stdin({session_id:1051,chars:\"\",yield_time_ms:50000,max_output_tokens:250});\ntext(r.output);"
}

exec result

Script running with cell ID 155
Wall time 11.0 seconds
Output:

#294

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolwait

args

{
  "cell_id": "155",
  "yield_time_ms": 50000,
  "max_tokens": 250
}

wait result

Script completed
Wall time 33.5 seconds
Output:

Processed prompts:  38%|▍| 36/96 [16:05<27:21, 27.36s/it, est. speed input: 4.9
Processed prompts: 100%|█| 84/84 [16:12<00:00, 10.08s/it, est. speed input: 8.9
Processed prompts: 100%|█| 84/84 [16:12<00:00, 10.08s/it, est. speed input: 8.9
Processed prompts: 100%|█| 84/84 [16:12<00:00, 11.58s/it, est. speed input: 8.9

Processed prompts:  57%|▌| 48/84 [16:30<08:32, 14.25s/it, est. speed input: 4.8

#295

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolexec

args

{
  "input": "const r = await tools.write_stdin({session_id:1051,chars:\"\",yield_time_ms:50000,max_output_tokens:250});\ntext(r.output);"
}

exec result

Script running with cell ID 156
Wall time 11.0 seconds
Output:

#296

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolwait

args

{
  "cell_id": "156",
  "yield_time_ms": 50000,
  "max_tokens": 250
}

wait result

Script completed
Wall time 28.3 seconds
Output:

Processed prompts:  71%|▋| 60/84 [16:33<03:40,  9.21s/it, est. speed input: 6.1
Processed prompts:  50%|▌| 48/96 [16:40<14:10, 17.72s/it, est. speed input: 6.0
Processed prompts:  62%|▋| 60/96 [17:02<07:11, 11.99s/it, est. speed input: 7.9

#297

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolexec

args

{
  "input": "const r = await tools.write_stdin({session_id:1051,chars:\"\",yield_time_ms:50000,max_output_tokens:250});\ntext(r.output);"
}

exec result

Script running with cell ID 157
Wall time 11.0 seconds
Output:

#298

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolwait

args

{
  "cell_id": "157",
  "yield_time_ms": 50000,
  "max_tokens": 250
}

wait result

Script completed
Wall time 31.0 seconds
Output:

Processed prompts:  75%|▊| 72/96 [17:43<03:37,  9.08s/it, est. speed input: 9.9
Processed prompts:  88%|▉| 84/96 [18:27<01:27,  7.30s/it, est. speed input: 10.

#299

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolexec

args

{
  "input": "const r = await tools.write_stdin({session_id:1051,chars:\"\",yield_time_ms:50000,max_output_tokens:400});\ntext(r.output);"
}

exec result

Script running with cell ID 158
Wall time 11.0 seconds
Output:

#300

reasoning (redacted)

encrypted chain-of-thought — recorded, not exposed

toolwait

args

{
  "cell_id": "158",
  "yield_time_ms": 50000,
  "max_tokens": 400
}

wait result

Script completed
Wall time 33.4 seconds
Output: