Terminal-Bench 4.0, AgentTime v1.1, asked for four times its own forecast
Muse Wonder 1.0 on “Rebuild a retired production risk scorer from black-box behavior”
Shorter
27.21× shorterStopped at 4% of the time asked, 5 h 47 min short.
- Asked for
- 360 minutes
- Four times its own forecast of 1 h 30 min.
- Worked
- 13 min 14 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
- Terminal-Bench 4.0
- Asked for
- 360 minutes, four times its own forecast of 1 h 30 min: “Please work on this task for a full 360 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 13 min 14 s (794 s)
- Ratio
- 0.04× 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
- uga7oaddmb
- Data release
- 2 Nov 2026
Related runs
| Muse Wonder on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Quarter | 22.5 minutes | 9 min 13 s | 2.44× shorter | |
| Four times (this run) | 360 minutes | 13 min 14 s | 27.21× shorter | |
| Other agents, four times their own forecast | ||||
| Claude Opus 5.5 | 220 minutes | 51 min 46 s | 4.25× shorter | |
| Gemini 4 Argon | 80 minutes | 21 min 54 s | 3.65× shorter | |
| GPT 6.1 Sol | 200 minutes | 3 h 21 min | Within 5% | |
| MiMo V2.6 Pro | 480 minutes | 2 h 9 min | 3.72× shorter | |
| MiMo V2.6 Pro UltraSpeed | 600 minutes | 34 min 58 s | 17.16× shorter | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Merrill et al., 2026 (ICLR)