Search-SWE
CPU TASK Valid result

Negative-Constraint Retrieval

GPT-6 Luna · Codex (v0.164.0) · reasoning effort max

Task score (%)0.0
Input Tokens11,078,228including 10,864,413 cached
Output Tokens166,738
Agent time40min
Cost$0.22
Recorded steps78

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
2026-10-24T03:26:23.880399Z
Coding finished
2026-10-24T04:06:40.810814Z
Input (includes cache)
11,078,228
Cached input (subset)
10,864,413
Cache writes (subset)
211,518
Output
166,738
Reasoning (reported separately; not added)
125,719
Recorded responses
75

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": 7200.0,
  "coding_commit": null,
  "effort": "max",
  "model": "openai/gpt-6-luna",
  "provider": "OpenRouter",
  "resources": null,
  "run_mode": "full",
  "task_commit": "bb34e915be5c2b2948146c2d7862cb58652d6300",
  "verifier_commit": "1537d01c332d4c69160e7106757ac32602ba79b3",
  "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
negative-constraint-retrieval-gpt-6-luna-max-81a2735fc844
Source trajectory SHA-256
f5141918439e852bce99bc021a1e91f3bf555a6c5c7cb37afe282d3fc5f51a3c
Final reward SHA-256
e51c8aa464a110f769a86cc20898314713b79b8d93a77a325a34684b994d6aaf