PPTArena, longest of three requests
GPT 6 Astra on “Arabic translation and layout”
Longer
1.48× longerWorked 48% past the request, 11 min 53 s over.
- Asked for
- 25 minutes
- Worked
- 36 min 53 s
- Grade
- Score 60 of 100
- Instructions 2 of 5, visuals 4 of 5. Graded during the run.
- How it ended
- Ended on its own
Every detail of this run
- Agent
- GPT 6 Astra, in Codex
- Benchmark
- PPTArena
- Asked for
- 25 minutes: “Please work on this task for a full 25 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 36 min 53 s (2,213 s)
- Ratio
- 1.48× the request
- How it ended
- Ended on its own
- Grade
- Instructions 2.00 of 5, visuals 4.00 of 5
- Two 0-to-5 judge scores, instruction following (IF) and visual quality (VQ), each the median of 3 judge samples. No official pass rule.
- Score
- 60 of 100
- Graded during the run. The paper's value.
- Wording
- The standard sentence, used from 4 September 2026.
- Run id
- 2xwcquej7h
- Data release
- 26 Sep 2026
Related runs
| Astra on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Shortest | 1.5 minutes | 4 min 41 s | 3.12× longer | |
| Middle | 6 minutes | 6 min 16 s | Within 5% | |
| Longest (this run) | 25 minutes | 36 min 53 s | 1.48× longer | |
| Other agents, longest request | ||||
| GPT 5.6 Sol | 25 minutes | 14 min 24 s | 1.74× shorter | |
| Claude Fable 5.1 | 25 minutes | 18 min 44 s | 1.33× shorter | |
See it on the chart. Transcripts are not published yet.