GPU TASK Valid result
Multilingual Vector Postprocessing
GPT-6 Luna · Codex · reasoning effort max
Task score (%)26.4
Input Tokens16,088,069including 15,888,705 cached
Output Tokens142,560
Agent time1h 4min
Cost$0.26
Recorded steps132
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
- 2026-10-23T22:58:34.116645Z
- Coding finished
- 2026-10-24T00:02:59.644423Z
- Input (includes cache)
- 16,088,069
- Cached input (subset)
- 15,888,705
- Cache writes (subset)
- 198,462
- Output
- 142,560
- Reasoning (reported separately; not added)
- 105,653
- Recorded responses
- 128
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": null,
"run_mode": "full_agent_and_verifier",
"task_commit": null,
"verifier_commit": "8287076ba397b9266c9a399a857a89ba4815e2cb",
"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
multilingual-vector-postprocessing-gpt-6-luna-max-bd8f388464da- Source trajectory SHA-256
19c67f6f75cff6e92d598afd0f65a4317feaeddf85908847c75018d5d156cbb2- Final reward SHA-256
b8df27dbcfccc449235527fbf191037737909fcf1a82507ecaf2f9fc9d54c1d4