AssistantBench, middle of three requests
GPT 5.6 Sol on “AssistantBench validation task 1”
Shorter
1.98× shorterStopped at 50% of the time asked, 2 min 29 s short.
- Asked for
- 5 minutes
- Worked
- 2 min 31 s
- Grade
- Did not pass
- Score 37.5 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
- 5 minutes: “Please work on this task for a full 5 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 2 min 31 s (151 s)
- Ratio
- 0.50× the request
- How it ended
- Ended on its own
- Grade
- Did not pass
- It did not match the reference answer exactly. The grader returned 0.375. Pass rule: Exact match with the reference answer.
- Score
- 37.5 of 100
- Graded afterwards. The paper's value.
- Wording
- The standard sentence, used from 4 September 2026.
- Run id
- obabvxte6i
- Data release
- 26 Sep 2026
Related runs
| Sol on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Shortest | 1.25 minutes | 58 s | 1.30× shorter | |
| Middle (this run) | 5 minutes | 2 min 31 s | 1.98× shorter | |
| Longest | 20 minutes | 17 min 5 s | 1.17× shorter | |
| Other agents, middle request | ||||
| GPT 6 Astra | 5 minutes | 5 min 44 s | 1.15× longer | |
| Claude Fable 5.1 | 5 minutes | 5 min 29 s | 1.10× longer | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Yoran et al., 2024 (EMNLP)