Humanity's Last Exam, AgentTime v1.1, asked for a quarter of its own forecast
Muse Wonder 1.0 on “Humanity's Last Exam item 39”
Longer
2.36× longerWorked 136% past the request, 20 s over.
- Asked for
- 15 seconds
- A quarter of its own forecast of 1 min.
- Worked
- 35 s
- Grade
- Not graded
- How it ended
- Ended on its own
Every detail of this run
- Agent
- Muse Wonder 1.0, in Muse Code
- Benchmark
- Humanity's Last Exam
- Asked for
- 15 seconds, a quarter of its own forecast of 1 min: “Please work on this task for a full 15 seconds. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 35 s (35 s)
- Ratio
- 2.36× the request
- How it ended
- Ended on its own
- Grade
- Not graded
- Clock tool
- The agent could call a wall-clock tool to read the current time.
- Wording
- The paper's standard sentence, sent in AgentTime v1.1.
- Run id
- mw2n3fvxd6
- Data release
- 2 Nov 2026
Related runs
| Muse Wonder on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Quarter (this run) | 15 seconds | 35 s | 2.36× longer | |
| Four times | 240 seconds | 2 min 17 s | 1.75× shorter | |
| Other agents, a quarter of their own forecast | ||||
| Claude Opus 5.5 | 22.5 seconds | 54 s | 2.39× longer | |
| Gemini 4 Argon | 6 seconds | 30 s | 4.93× longer | |
| GPT 6.1 Sol | 22.5 seconds | 37 s | 1.65× longer | |
| MiMo V2.6 Pro | 30 seconds | 3 min 20 s | 6.67× longer | |
| MiMo V2.6 Pro UltraSpeed | 45 seconds | 44 s | Within 5% | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Phan et al., 2025