Sakana ALE-Bench, AgentTime v1.1, asked for four times its own forecast
Muse Wonder 1.0 on “AtCoder Heuristic Contest 11”
Shorter
153.96× shorterStopped at 1% of the time asked, 25 h 8 min short.
- Asked for
- 25.3 hours
- Four times its own forecast of 6 h 20 min.
- Worked
- 9 min 52 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
- Sakana ALE-Bench
- Asked for
- 25.3 hours, four times its own forecast of 6 h 20 min: “Please work on this task for a full 25.3 hours. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 9 min 52 s (592 s)
- Ratio
- 0.01× 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
- fgcw4k4jwf
- Data release
- 2 Nov 2026
Related runs
| Muse Wonder on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Quarter | 1.58 hours | 7 min 2 s | 13.49× shorter | |
| Four times (this run) | 25.3 hours | 9 min 52 s | 153.96× shorter | |
| Other agents, four times their own forecast | ||||
| Claude Opus 5.5 | 400 minutes | 5 h 18 min | 1.26× shorter | |
| Gemini 4 Argon | 180 minutes | 1 h 13 min | 2.47× shorter | |
| GPT 6.1 Sol | 180 minutes | 3 h 1 min | Within 5% | |
| MiMo V2.6 Pro | 24 hours | 2 h 28 min | 9.73× shorter | |
| MiMo V2.6 Pro UltraSpeed | 22 hours | 3 h 35 min | 6.13× shorter | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Imajuku et al., 2025