TUA-Bench, AgentTime v1.1, asked for four times its own forecast
Claude Opus 5.5 on “Extract gym auditorium”
Shorter
3.54× shorterStopped at 28% of the time asked, 57 min 25 s short.
- Asked for
- 80 minutes
- Four times its own forecast of 20 min.
- Worked
- 22 min 35 s
- Grade
- Not graded
- How it ended
- Ended on its own
Every detail of this run
- Agent
- Claude Opus 5.5, in Claude Code
- Benchmark
- TUA-Bench
- Asked for
- 80 minutes, four times its own forecast of 20 min: “Please work on this task for a full 80 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 22 min 35 s (1,355 s)
- Ratio
- 0.28× 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
- ueexv3z5uh
- Data release
- 2 Nov 2026
Related runs
| Opus 5.5 on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Quarter | 5 minutes | 16 min 3 s | 3.21× longer | |
| Four times (this run) | 80 minutes | 22 min 35 s | 3.54× shorter | |
| Other agents, four times their own forecast | ||||
| Gemini 4 Argon | 48 minutes | 48 min 31 s | Within 5% | |
| GPT 6.1 Sol | 32 minutes | 32 min 23 s | Within 5% | |
| MiMo V2.6 Pro | 200 minutes | 21 min 32 s | 9.29× shorter | |
| MiMo V2.6 Pro UltraSpeed | 300 minutes | 12 min 33 s | 23.90× shorter | |
| Muse Wonder 1.0 | 240 minutes | 2 min 19 s | 103.90× shorter | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Chen et al., 2026