ProgramBench, middle of three requests

GPT 6 Astra on “facebook/zstd”

Within 5%

Worked 101% of the time asked, within 5% of it.

Asked

Worked

Safety cutoff 5 h, off the scale

Every detail of this run

Agent: GPT 6 Astra, in Codex
Benchmark: ProgramBench
Task: facebook/zstd
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: 40 min 35 s (2,435 s)
Ratio: 1.01× the request
How it ended: Ending not labelled yet
Grade: Did not pass; It did not pass every hidden test. The grader returned 0.783. Pass rule: Resolved: every hidden test passes.
Score: 78.3 of 100; Graded afterwards. The paper's value.
Wording: The standard sentence, used from 4 September 2026.
Run id: x5n4njajjh
Data release: 26 Sep 2026

The other requests for GPT 6 Astra on this task, and the other agents at this request
Astra on this task	Asked	Worked	Result
Shortest	10 minutes	10 min 12 s	Within 5%
Middle (this run)	40 minutes	40 min 35 s	Within 5%
Longest	150 minutes	2 h 32 min	Within 5%
Other agents, middle request
GPT 5.6 Sol	40 minutes	40 min 12 s	Within 5%
Claude Fable 5.1	40 minutes	20 min 18 s	1.97× shorter

See it on the chart. Transcripts are not published yet.

Read before quoting. How this run ended has not been labelled yet. It is timed like every other run and counted in every number. Details

Data and citation