AssistantBench, longest of three requests
GPT 5.6 Sol on “AssistantBench validation task 3”
Shorter
1.07× shorterStopped at 93% of the time asked, 1 min 22 s short.
- Asked for
- 20 minutes
- Worked
- 18 min 38 s
- Grade
- Did not pass
- Score 0 of 100. Graded afterwards.
- How it ended
- Ended on its own
Every detail of this run
- Agent
- GPT 5.6 Sol, in Codex
- Benchmark
- AssistantBench
- Asked for
- 20 minutes: “Please work on this task for a full 20 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 18 min 38 s (1,118 s)
- Ratio
- 0.93× the request
- How it ended
- Ended on its own
- Grade
- Did not pass
- It did not match the reference answer exactly. The grader returned 0. Pass rule: Exact match with the reference answer.
- Score
- 0 of 100
- Graded afterwards. The paper's value.
- Wording
- The standard sentence, used from 4 September 2026.
- Run id
- 5p3bacjjyb
- Data release
- 26 Sep 2026
Related runs
| Sol on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Shortest | 1.25 minutes | 6 min 12 s | 4.96× longer | |
| Middle | 5 minutes | 3 min 41 s | 1.36× shorter | |
| Longest (this run) | 20 minutes | 18 min 38 s | 1.07× shorter | |
| Other agents, longest request | ||||
| GPT 6 Astra | 20 minutes | 5 min 51 s | 3.42× shorter | |
| Claude Fable 5.1 | 20 minutes | 8 min 6 s | 2.47× shorter | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Yoran et al., 2024 (EMNLP)