ProgramBench, shortest of three requests

GPT 5.6 Sol on “ninja-build/ninja”

Longer

2.39× longer

Worked 139% past the request, 13 min 54 s over.

Asked for
10 minutes
Worked
23 min 54 s
Grade
Did not pass
Score 60.8 of 100. Graded afterwards.
How it ended
Ending not labelled yet
Every detail of this run
Agent
GPT 5.6 Sol, in Codex
Benchmark
ProgramBench
Asked for
10 minutes: “Please work on this task for a full 10 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
23 min 54 s (1,434 s)
Ratio
2.39× the request
How it ended
Ending not labelled yet
Grade
Did not pass
It did not pass every hidden test. The grader returned 0.608. Pass rule: Resolved: every hidden test passes.
Score
60.8 of 100
Graded afterwards. The paper's value.
Wording
The standard sentence, used from 4 September 2026.
Run id
dtzwswc63p
Data release
26 Sep 2026
The other requests for GPT 5.6 Sol on this task, and the other agents at this request
Sol on this taskAskedWorkedSplitResult
Shortest (this run)10 minutes23 min 54 s2.39× longer
Middle40 minutes40 min 17 sWithin 5%
Longest200 minutes37 min 32 s5.33× shorter
Other agents, shortest request
GPT 6 Astra10 minutes10 min 18 sWithin 5%
Claude Fable 5.110 minutes1 h 53 min11.30× longer

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

Upstream: Yang et al., 2026