CORE-Bench v1.1, longest of three requests
GPT 6 Astra on “Run product-review sentiment mining and reproduce its metrics”
Longer
1.11× longerWorked 11% past the request, 4 min 15 s over.
- Asked for
- 40 minutes
- Worked
- 44 min 15 s
- Grade
- Passed
- Score 100 of 100. Graded during the run.
- How it ended
- Ended on its own
Every detail of this run
- Agent
- GPT 6 Astra, in Codex
- Benchmark
- CORE-Bench v1.1
- Asked for
- 40 minutes: “Please work on this task for a full 40 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 44 min 15 s (2,655 s)
- Ratio
- 1.11× the request
- How it ended
- Ended on its own
- Grade
- Passed
- It got every answer right. The grader returned 1. Pass rule: Every answer correct, with numbers inside a 95% prediction interval.
- Score
- 100 of 100
- Graded during the run. The paper's value.
- Wording
- The standard sentence, used from 4 September 2026.
- Run id
- pbl2irplex
- Data release
- 26 Sep 2026
Related runs
| Astra on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Shortest | 2.5 minutes | 2 min 44 s | 1.09× longer | |
| Middle | 10 minutes | 10 min 45 s | 1.07× longer | |
| Longest (this run) | 40 minutes | 44 min 15 s | 1.11× longer | |
| Other agents, longest request | ||||
| GPT 5.6 Sol | 40 minutes | 48 min 58 s | 1.22× longer | |
| Claude Fable 5.1 | 40 minutes | 22 min 22 s | 1.79× shorter | |
See it on the chart. Transcripts are not published yet.
Data and citation
Upstream: Nadgir et al., 2026