PaperBench, longest of three requests
GPT 5.6 Sol on “FRE”
Within 5%
Within 5%Worked 100% of the time asked, within 5% of it.
- Asked for
- 150 minutes
- Worked
- 2 h 31 min
- Grade
- Score 17.2 of 100
- Judged afterwards.
- How it ended
- Ended on its own
Every detail of this run
- Agent
- GPT 5.6 Sol, in Codex
- Benchmark
- PaperBench
- Task
- FRE
- Asked for
- 150 minutes: “Please work on this task for a full 150 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 2 h 31 min (9,034 s)
- Ratio
- 1.00× the request
- How it ended
- Ended on its own
- Grade
- 0.172
- Replication score: a weighted rubric of pass/fail leaf checks, 0 to 1, after a full reproduction attempt. No official pass rule.
- Score
- 17.2 of 100
- Judged afterwards. The paper's value.
- Wording
- The standard sentence, used from 4 September 2026.
- Run id
- qaqpskz5ct
- Data release
- 26 Sep 2026
Related runs
| Sol on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Shortest | 10 minutes | 6 min 27 s | 1.55× shorter | |
| Middle | 40 minutes | 27 min 36 s | 1.45× shorter | |
| Longest (this run) | 150 minutes | 2 h 31 min | Within 5% | |
| Other agents, longest request | ||||
| GPT 6 Astra | 150 minutes | 2 h 31 min | Within 5% | |
| Claude Fable 5.1 | 150 minutes | 42 min 3 s | 3.57× shorter | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Starace et al., 2025