Terminal-Bench 4.0, AgentTime v1.1, asked for four times its own forecast
Muse Wonder 1.0 on “Recalculate SA-CCR exposure, RWA, capital, and workbook evidence”
Shorter
128.00× shorterStopped at 1% of the time asked, 10 h 35 min short.
- Asked for
- 640 minutes
- Four times its own forecast of 2 h 40 min.
- Worked
- 5 min
- 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
- 640 minutes, four times its own forecast of 2 h 40 min: “Please work on this task for a full 640 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 5 min (300 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
- lvhi5wk5rv
- Data release
- 2 Nov 2026
Related runs
| Muse Wonder on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Quarter | 40 minutes | 7 min 59 s | 5.01× shorter | |
| Four times (this run) | 640 minutes | 5 min | 128.00× shorter | |
| Other agents, four times their own forecast | ||||
| Claude Opus 5.5 | 120 minutes | 34 min 43 s | 3.46× shorter | |
| Gemini 4 Argon | 72 minutes | 23 min 43 s | 3.04× shorter | |
| GPT 6.1 Sol | 88 minutes | 1 h 29 min | Within 5% | |
| MiMo V2.6 Pro | 720 minutes | 22 min 58 s | 31.36× shorter | |
| MiMo V2.6 Pro UltraSpeed | 380 minutes | 8 min 1 s | 47.38× shorter | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Merrill et al., 2026 (ICLR)