PostTrainBench v1.1, AgentTime v1.1, asked for a quarter of its runtime with no time asked
Gemini 4 Argon on “Post-train Qwen3-1.7B-Base for GSM8K”
Longer
1.27× longerWorked 27% past the request, 21 min 57 s over.
- Asked for
- 1.35 hours
- A quarter of its runtime with no time asked (5 h 25 min).
- Worked
- 1 h 43 min
- Grade
- Not graded
- How it ended
- Ended on its own
Every detail of this run
- Agent
- Gemini 4 Argon, in Antigravity CLI
- Benchmark
- PostTrainBench v1.1
- Asked for
- 1.35 hours, a quarter of its runtime with no time asked (5 h 25 min): “Please work on this task for a full 1.35 hours. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 1 h 43 min (6,177 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
- yvwjsj7e4x
- Data release
- 2 Nov 2026
Related runs
| Argon on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Quarter (this run) | 1.35 hours | 1 h 43 min | 1.27× longer | |
| Four times | 21.7 hours | 4 h 8 min | 5.25× shorter | |
| Other agents, a quarter of their own runtime | ||||
| Claude Opus 5.5 | 4.85 hours | 4 h 13 min | 1.15× shorter | |
| GPT 6.1 Sol | 47.6 minutes | 48 min 21 s | Within 5% | |
| MiMo V2.6 Pro | 50.2 minutes | 2 h 53 min | 3.45× longer | |
| MiMo V2.6 Pro UltraSpeed | 15.5 seconds | 2 min 59 s | 11.57× longer | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Rank et al., 2026 (ICML)