Terminal-Bench 4.0, longest of three requests
GPT 6 Astra on “Rebuild a retired production risk scorer from black-box behavior”
Within 5%
Within 5%Worked 101% of the time asked, within 5% of it.
- Asked for
- 100 minutes
- Worked
- 1 h 41 min
- Grade
- Did not pass
- Score 0 of 100. Graded afterwards.
- How it ended
- Ended on its own
Every detail of this run
- Agent
- GPT 6 Astra, in Codex
- Benchmark
- Terminal-Bench 4.0
- Asked for
- 100 minutes: “Please work on this task for a full 100 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 1 h 41 min (6,039 s)
- Ratio
- 1.01× the request
- How it ended
- Ended on its own
- Grade
- Did not pass
- It did not pass the task's tests. The grader returned 0. Pass rule: The task's tests pass.
- Score
- 0 of 100
- Graded afterwards. The paper's value.
- Wording
- The standard sentence, used from 4 September 2026.
- Run id
- wrgeioo36x
- Data release
- 26 Sep 2026
Related runs
| Astra on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Shortest | 6 minutes | 6 min 12 s | Within 5% | |
| Middle | 25 minutes | 25 min 15 s | Within 5% | |
| Longest (this run) | 100 minutes | 1 h 41 min | Within 5% | |
| Other agents, longest request | ||||
| GPT 5.6 Sol | 100 minutes | 1 h 41 min | Within 5% | |
| Claude Fable 5.1 | 100 minutes | 22 min 46 s | 4.39× shorter | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Merrill et al., 2026 (ICLR)