CORE-Bench v1.1, middle of three requests
GPT 6 Astra on “Train and evaluate end-to-end neural-network feature selection on MNIST”
Within 5%
Within 5%Worked 102% of the time asked, within 5% of it.
- Asked for
- 50 minutes
- Worked
- 50 min 48 s
- Grade
- Did not pass
- Score 0 of 100. Graded afterwards.
- How it ended
- Stopped by an error
Every detail of this run
- Agent
- GPT 6 Astra, in Codex
- Benchmark
- CORE-Bench v1.1
- Asked for
- 50 minutes: “Please work on this task for a full 50 minutes. Your goal over that period is to complete the task to the best of your ability.”
- Worked
- 50 min 48 s (3,048 s)
- Ratio
- 1.02× the request
- How it ended
- Stopped by an error
- Grade
- Did not pass
- It did not get every answer right. The grader returned 0. Pass rule: Every answer correct, with numbers inside a 95% prediction interval.
- Score
- 0 of 100
- Graded afterwards. The paper's value.
- Wording
- The standard sentence, used from 4 September 2026.
- Run id
- 62skyfzgno
- Data release
- 26 Sep 2026
Related runs
| Astra on this task | Asked | Worked | Split | Result |
|---|---|---|---|---|
| Shortest | 12.5 minutes | 13 min 8 s | 1.05× longer | |
| Middle (this run) | 50 minutes | 50 min 48 s | Within 5% | |
| Longest | 200 minutes | 3 h 21 min | Within 5% | |
| Other agents, middle request | ||||
| GPT 5.6 Sol | 50 minutes | 3 h 28 min | 4.16× longer | |
| Claude Fable 5.1 | 50 minutes | 6 h 40 min | 8.00× longer | |
See it on the chart. Transcripts are not published yet.
Read before quoting. This run was stopped by an error, so its time is a lower bound: the agent might have kept going. Details
Data and citation
Upstream: Nadgir et al., 2026