Agents' Last Exam, longest of three requests
GPT 5.6 Sol on “Build a clean retail SQLite warehouse from messy inputs”
Longer
1.08× longerWorked 8% past the request, 6 min 8 s over.
- Asked for
- 80 minutes
- Worked
- 1 h 26 min
- Grade
- Did not pass
- Score 83.3 of 100. Graded afterwards.
- How it ended
- Ended on its own
Every detail of this run
- Agent
- GPT 5.6 Sol, in Codex
- Benchmark
- Agents' Last Exam
- Asked for
- 80 minutes: “Please work on this task for a full 80 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 1 h 26 min (5,168 s)
- Ratio
- 1.08× 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 0.833. Pass rule: Full pass: a checker score of 1.
- Score
- 83.3 of 100
- Graded afterwards. The paper's value.
- Wording
- The standard sentence, used from 4 September 2026.
- Run id
- jjs434cyom
- Data release
- 26 Sep 2026
Related runs
| Sol on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Shortest | 5 minutes | 6 min 9 s | 1.23× longer | |
| Middle | 20 minutes | 20 min 17 s | Within 5% | |
| Longest (this run) | 80 minutes | 1 h 26 min | 1.08× longer | |
| Other agents, longest request | ||||
| GPT 6 Astra | 80 minutes | 1 h 20 min | Within 5% | |
| Claude Fable 5.1 | 80 minutes | 22 min 37 s | 3.54× shorter | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Sun et al., 2026