Agents' Last Exam, longest of three requests
Claude Fable 5.1 on “Redesign and validate a Flowable supply-disruption workflow”
Shorter
1.87× shorterStopped at 53% of the time asked, 58 min 13 s short.
- Asked for
- 125 minutes
- Worked
- 1 h 7 min
- Grade
- Did not pass
- Score 89.4 of 100. Graded afterwards.
- How it ended
- Ended on its own
Every detail of this run
- Agent
- Claude Fable 5.1, in Claude Code
- Benchmark
- Agents' Last Exam
- Asked for
- 125 minutes: “Please work on this task for a full 125 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 1 h 7 min (4,007 s)
- Ratio
- 0.53× the request
- How it ended
- Ended on its own
- Grade
- Did not pass
- It did not pass the task's checker in full. The grader returned 0.894. Pass rule: Full pass: a checker score of 1.
- Score
- 89.4 of 100
- Graded afterwards. The paper's value.
- Wording
- The standard sentence, used from 4 September 2026.
- Run id
- wha2w24wc4
- Data release
- 26 Sep 2026
Related runs
| Fable on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Shortest | 8 minutes | 34 min | 4.25× longer | |
| Middle | 30 minutes | 1 h 2 min | 2.07× longer | |
| Longest (this run) | 125 minutes | 1 h 7 min | 1.87× shorter | |
| Other agents, longest request | ||||
| GPT 6 Astra | 125 minutes | 2 h 7 min | Within 5% | |
| GPT 5.6 Sol | 125 minutes | 2 h 6 min | Within 5% | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Sun et al., 2026