WildClawBench, middle of three requests
GPT 5.6 Sol on “Cut all three first-half Ferran Torres goals into a sub-30-second reel”
Longer
2.07× longerWorked 107% past the request, 10 min 41 s over.
- Asked for
- 10 minutes
- Worked
- 20 min 41 s
- Grade
- Score 30 of 100
- Graded during the run.
- How it ended
- Ended on its own
Every detail of this run
- Agent
- GPT 5.6 Sol, in Codex
- Benchmark
- WildClawBench
- Asked for
- 10 minutes: “Please work on this task for a full 10 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 20 min 41 s (1,241 s)
- Ratio
- 2.07× the request
- How it ended
- Ended on its own
- Grade
- 0.3
- Task grader score, 0 to 1 (a model judge only where the task calls for one). No official pass rule.
- Score
- 30 of 100
- Graded during the run. The paper's value.
- Wording
- The standard sentence, used from 4 September 2026.
- Run id
- 3vci6rohv5
- Data release
- 26 Sep 2026
Related runs
| Sol on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Shortest | 2.5 minutes | 18 min 57 s | 7.58× longer | |
| Middle (this run) | 10 minutes | 20 min 41 s | 2.07× longer | |
| Longest | 40 minutes | 40 min 52 s | Within 5% | |
| Other agents, middle request | ||||
| GPT 6 Astra | 10 minutes | 10 min 22 s | Within 5% | |
| Claude Fable 5.1 | 10 minutes | 34 min 2 s | 3.40× longer | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Ding et al., 2026