ProgramBench, AgentTime v1.1, asked for a quarter of its own forecast
Gemini 4 Argon on “rvben/rumdl”
Longer
1.62× longerWorked 62% past the request, 17 min 48 s over.
- Asked for
- 28.75 minutes
- A quarter of its own forecast of 1 h 55 min.
- Worked
- 46 min 33 s
- Grade
- Not graded
- How it ended
- Ended on its own
Every detail of this run
- Agent
- Gemini 4 Argon, in Antigravity CLI
- Benchmark
- ProgramBench
- Task
- rvben/rumdl
- Asked for
- 28.75 minutes, a quarter of its own forecast of 1 h 55 min: “Please work on this task for a full 28.75 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 46 min 33 s (2,793 s)
- Ratio
- 1.62× 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
- dtmbsarohc
- Data release
- 2 Nov 2026
Related runs
| Argon on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Quarter (this run) | 28.75 minutes | 46 min 33 s | 1.62× longer | |
| Four times | 460 minutes | 50 min 13 s | 9.16× shorter | |
| Other agents, a quarter of their own forecast | ||||
| Claude Opus 5.5 | 1 hour | 54 min 56 s | 1.09× shorter | |
| GPT 6.1 Sol | 30 minutes | 30 min 23 s | Within 5% | |
| MiMo V2.6 Pro | 5 hours | 51 min 44 s | 5.80× shorter | |
| MiMo V2.6 Pro UltraSpeed | 6.25 hours | 1 h 54 min | 3.30× shorter | |
| Muse Wonder 1.0 | 4 hours | 12 min 18 s | 19.51× shorter | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Yang et al., 2026