Terminal-Bench 4.0, AgentTime v1.1, asked for a quarter of its own forecast
GPT 6.1 Sol on “Reconstruct a 3D corner tie from a 2D schematic”
Longer
1.07× longerWorked 7% past the request, 15 s over.
- Asked for
- 3.5 minutes
- A quarter of its own forecast of 14 min.
- Worked
- 3 min 45 s
- Grade
- Not graded
- How it ended
- Ended on its own
Every detail of this run
- Agent
- GPT 6.1 Sol, in Codex
- Benchmark
- Terminal-Bench 4.0
- Asked for
- 3.5 minutes, a quarter of its own forecast of 14 min: “Please work on this task for a full 3.5 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 3 min 45 s (225 s)
- Ratio
- 1.07× 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
- bc4vtnj3am
- Data release
- 2 Nov 2026
Related runs
| Sol 6.1 on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Quarter (this run) | 3.5 minutes | 3 min 45 s | 1.07× longer | |
| Four times | 56 minutes | 56 min 58 s | Within 5% | |
| Other agents, a quarter of their own forecast | ||||
| Claude Opus 5.5 | 3.75 minutes | 2 min 50 s | 1.33× shorter | |
| Gemini 4 Argon | 4.5 minutes | 14 min 23 s | 3.20× longer | |
| MiMo V2.6 Pro | 18.75 minutes | 32 min 50 s | 1.75× longer | |
| MiMo V2.6 Pro UltraSpeed | 12.5 minutes | 14 min 2 s | 1.12× longer | |
| Muse Wonder 1.0 | 8 minutes | 10 min 23 s | 1.30× longer | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Merrill et al., 2026 (ICLR)