Sakana ALE-Bench, longest of three requests

GPT 5.6 Sol on “AtCoder Heuristic Contest 9”

Within 5%

Within 5%

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

Asked for
150 minutes
Worked
2 h 32 min
Grade
Performance 3482
Graded during the run.
How it ended
Ended on its own
Every detail of this run
Agent
GPT 5.6 Sol, in Codex
Asked for
150 minutes: “Please work on this task for a full 150 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
2 h 32 min (9,127 s)
Ratio
1.01× the request
How it ended
Ended on its own
Grade
7,718,070,185
Raw score on the private test cases, in each problem's own units; also reports rank against human contest entrants. No official pass rule.
Score
Performance 3482
Graded during the run. The paper's value.
Wording
The standard sentence, used from 4 September 2026.
Run id
mpixkekwoc
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
Shortest10 minutes11 min1.10× longer
Middle40 minutes40 min 23 sWithin 5%
Longest (this run)150 minutes2 h 32 minWithin 5%
Other agents, longest request
GPT 6 Astra150 minutes2 h 31 minWithin 5%
Claude Fable 5.1150 minutes1 h 22 min1.82× shorter

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

Data and citation

Upstream: Imajuku et al., 2025