GPU TASK Valid result
Layout Page Search
GPT-6 Luna · Codex · reasoning effort max
Task score (%)18.8
Input Tokens14,568,292including 14,393,597 cached
Output Tokens104,329
Agent time1h 4min
Cost$0.22
Recorded steps141
Cost is estimated from recorded agent tokens at official API rates in USD (2026-11-01; DeepSeek off-peak pricing).
Loading trajectory…
Select this tab to browse the saved files.
Coding time & token usage
- Original agent completion
- normal
- Coding started
- 2026-10-25T12:45:20.232671Z
- Coding finished
- 2026-10-25T13:49:28.237794Z
- Input (includes cache)
- 14,568,292
- Cached input (subset)
- 14,393,597
- Cache writes (subset)
- 174,290
- Output
- 104,329
- Reasoning (reported separately; not added)
- 76,875
- Recorded responses
- 135
Input includes cached input; output is counted once. Auxiliary reasoning counts are shown as recorded. Regrading does not repeat coding usage.
Select this tab to browse the saved files.
Configuration
Evaluation settings summary
{
"agent": "Codex",
"agent_timeout_sec": 21600.0,
"coding_commit": null,
"effort": "max",
"model": "openai/gpt-6-luna",
"provider": "OpenRouter",
"resources": {
"cpus": 8,
"gpus": 1,
"memory_mb": 32768,
"storage_mb": 102400
},
"run_mode": "official_harbor_full",
"task_commit": "052045836520335c1caa5e4a06ed887faacef55f",
"verifier_commit": "052045836520335c1caa5e4a06ed887faacef55f",
"version": "0.164.0"
}Select this tab to browse the saved files.
Resolved configuration
Select this tab to browse the saved files.
Trial log
Select this tab to browse the saved files.
Exceptions
Select this tab to browse the saved files.
Select this tab to browse the saved files.
Trajectory downloads & provenance
Public text redacts private values and normalizes host paths. Source hashes identify the original records.
- Trial ID
layout-page-search-gpt-6-luna-max-e5215999abc9- Source trajectory SHA-256
04977158fc8df3975e4246b36602c092e22aca1a80d2713eae39e9836f07274d- Final reward SHA-256
a627b901722a6d12fac5addbefae12f1f31cd1004e3a550036627d1d6d16492d