Search-SWE
CPU TASK Valid result

Join-aware Table Retrieval

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

Task score (%)40.0
Input Tokens10,151,616including 9,922,632 cached
Output Tokens150,772
Agent time37min
Cost$0.20
Recorded steps86

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
completed
Coding started
2026-10-23T20:10:20.739062Z
Coding finished
2026-10-23T20:48:11.617813Z
Input (includes cache)
10,151,616
Cached input (subset)
9,922,632
Cache writes (subset)
202,042
Output
150,772
Reasoning (reported separately; not added)
100,179
Recorded responses
83

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": null,
  "coding_commit": null,
  "effort": "max",
  "model": "openai/gpt-6-luna",
  "provider": "OpenRouter",
  "resources": null,
  "run_mode": null,
  "task_commit": null,
  "verifier_commit": "bb34e915be5c2b2948146c2d7862cb58652d6300",
  "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
join-aware-table-retrieval-gpt-6-luna-max-eea6eb0d8435
Source trajectory SHA-256
c4876fa0ac4f9171fb31b1d8315d9a897b846dff7c8fa1351368ce92a5730fac
Final reward SHA-256
14e4ebf72b95cfb2a439b76633b406ede9071ab04ffccd9baaa5a4dbd1353ed6