OSWorld 2.0, AgentTime v1.1, asked for a quarter of its own forecast
GPT 6.1 Sol on “Invoices, photos, and email”
Longer
1.27× longerWorked 27% past the request, 3 min 1 s over.
- Asked for
- 11.25 minutes
- A quarter of its own forecast of 45 min.
- Worked
- 14 min 16 s
- Grade
- Not graded
- How it ended
- Ended on its own
Every detail of this run
- Agent
- GPT 6.1 Sol, in Codex
- Benchmark
- OSWorld 2.0
- Asked for
- 11.25 minutes, a quarter of its own forecast of 45 min: “Please work on this task for a full 11.25 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 14 min 16 s (856 s)
- Ratio
- 1.27× 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
- torqwzf45h
- Data release
- 2 Nov 2026
Related runs
| Sol 6.1 on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Quarter (this run) | 11.25 minutes | 14 min 16 s | 1.27× longer | |
| Four times | 180 minutes | 3 h 4 min | Within 5% | |
| Other agents, a quarter of their own forecast | ||||
| Claude Opus 5.5 | 10 minutes | 56 min 23 s | 5.64× longer | |
| Gemini 4 Argon | 5 minutes | 1 h 40 min | 19.92× longer | |
| MiMo V2.6 Pro | 9.5 minutes | 10 min 45 s | 1.13× longer | |
| MiMo V2.6 Pro UltraSpeed | 17.5 minutes | 1 h 48 min | 6.16× longer | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Yuan et al., 2026