AR to diffusion conversion

After grading, a reviewer model reads the submission's files and the agent's trajectory and checks each of the task's rules. A failed check makes the score 0. Step numbers link into the trajectory.

LLM adversarial review is not proof of arbitrary program behavior. Host storage and task-specific deterministic checks remain required.

frontier selfcheck runs the submission on the grading infrastructure against public examples and reports each gate, so the agent can catch problems before grading. Each run gets four. This agent used 3 of 4.

1. 1
   failed2h 35m into the runtook 4m 25s

   startup 40.2 s of 300 s; 0/1 gates passed; failed single-stream speed: 1 of 8 measured prompts, sent one at a time after an untimed warm-up (max_tokens 1024, temperature 1), each reply's reasoning and answer tokens counted with the pinned openbmb/MiniCPM5-2B tokenizer over end-to-end request

   3 gates

   • ✓review prerequisites
   • ✓startup · GET /health answered 200 after 40.2 s
   • ✗single-stream speed · 1 of 8 measured prompts, sent one at a time after an untimed warm-up (max_tokens 1024, temperature 1), each reply's reasoning and answer tokens counted with the pinned openbmb/MiniCPM5-2B tokenizer over end-to-end request time (sandbox clock), a miss re-sent once; slowest 424 tokens/s (1021 tokens in 2.41 s); per prompt: 424. Public prompts: 7 GPQA and 2 hard LiveCodeBench; grading draws 4 GPQA, 3 level-5 MATH-500 (held out) and 2 hard LiveCodeBench, so its replies differ. Grading stops here with score 0.
2. 2
   failed5h 17m into the runtook 7m 34s

   failed startup: Server did not become healthy in 300s (last /health probe: RuntimeError: Submission HTTP request failed: ConnectionRefusedError: [Errno 111] Connection refused); end of its log: [serve.sh] python: /opt/venv/bin/python (Python 3.12.3) [serve.sh] installing vendored wheels + t

   2 gates

   • ✓review prerequisites
   • ✗startup · Server did not become healthy in 300s (last /health probe: RuntimeError: Submission HTTP request failed: ConnectionRefusedError: [Errno 111] Connection refused); end of its log: [serve.sh] python: /opt/venv/bin/python (Python 3.12.3) [serve.sh] installing vendored wheels + typer==0.27.2 + typing-inspection==0.4.4 + uvicorn==0.54.0 [serve.sh] wheels installed in 12s /submission/serve.sh: line 48: BDLM_EXTRA_ARGS: unbound variable
3. 3
   passed5h 25m into the runtook 3m 57s

   startup 48.7 s of 300 s; 1/1 gates passed; The suite's 32 concurrent requests per copy, up to 3600 s each, were not sent; one request at its settings was.

   5 gates

   • ✓review prerequisites
   • ✓startup · GET /health answered 200 after 48.7 s
   • ✓single-stream speed · 8 of 8 measured prompts, sent one at a time after an untimed warm-up (max_tokens 1024, temperature 1), each reply's reasoning and answer tokens counted with the pinned openbmb/MiniCPM5-2B tokenizer over end-to-end request time (sandbox clock), a miss re-sent once; slowest 564 tokens/s (1024 tokens in 1.81 s); per prompt: 664, 723, 639, 688, 812, 564, 698, 689. Public prompts: 7 GPQA and 2 hard LiveCodeBench; grading draws 4 GPQA, 3 level-5 MATH-500 (held out) and 2 hard LiveCodeBench, so its replies differ.
   • ✓GPQA request as graded (self-check cap 291 s; grading allows 3600 s) · {"finish_reason": "stop", "answer": "C", "content_characters": 1098, "reasoning_characters": 34681, "content_end": "system, which is now reduced. Therefore, the endpoint decreases, even though the spectrum remains continuous.\n\nANSWER: C"}
   • ✓healthy after requests

The agent's last message before its session ended.