Agents' Last Exam, AgentTime v1.1, asked for four times its own forecast
MiMo V2.6 Pro on “Build a clean retail SQLite warehouse from messy inputs”
Shorter
33.29× shorterStopped at 3% of the time asked, 3 h 53 min short.
- Asked for
- 240 minutes
- Four times its own forecast of 1 h.
- Worked
- 7 min 13 s
- Grade
- Not graded
- How it ended
- Ended on its own
Every detail of this run
- Agent
- MiMo V2.6 Pro, in MiMo Code
- Benchmark
- Agents' Last Exam
- Asked for
- 240 minutes, four times its own forecast of 1 h: “Please work on this task for a full 240 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 7 min 13 s (433 s)
- Ratio
- 0.03× 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
- k3hiog2r7z
- Data release
- 2 Nov 2026
Related runs
| MiMo Pro on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Quarter | 15 minutes | 6 min 41 s | 2.24× shorter | |
| Four times (this run) | 240 minutes | 7 min 13 s | 33.29× shorter | |
| Other agents, four times their own forecast | ||||
| Claude Opus 5.5 | 128 minutes | 39 min 37 s | 3.23× shorter | |
| Gemini 4 Argon | 56 minutes | 14 min 58 s | 3.74× shorter | |
| GPT 6.1 Sol | 64 minutes | 1 h 4 min | Within 5% | |
| MiMo V2.6 Pro UltraSpeed | 248 minutes | 4 min 18 s | 57.67× shorter | |
| Muse Wonder 1.0 | 200 minutes | 1 min 8 s | 176.99× shorter | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Sun et al., 2026