WildClawBench, shortest of three requests

GPT 5.6 Sol on “Recover arXiv titles, figure counts, and BibTeX from 21 local PDFs”

Longer

2.63× longer

Worked 163% past the request, 4 min 5 s over.

Asked

Worked

Safety cutoff 1 h 20 min, off the scale

Every detail of this run

Agent: GPT 5.6 Sol, in Codex
Benchmark: WildClawBench
Task: Recover arXiv titles, figure counts, and BibTeX from 21 local PDFs
Asked for: 2.5 minutes: “Please work on this task for a full 2.5 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked: 6 min 35 s (395 s)
Ratio: 2.63× the request
How it ended: Ended on its own
Grade: 0.5; Task grader score, 0 to 1 (a model judge only where the task calls for one). No official pass rule.
Score: 50.1 of 100; Graded during the run. The paper's value.
Wording: The standard sentence, used from 4 September 2026.
Run id: cmvogtkogs
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 task	Asked	Worked	Result
Shortest (this run)	2.5 minutes	6 min 35 s	2.63× longer
Middle	10 minutes	8 min 23 s	1.19× shorter
Longest	40 minutes	41 min 36 s	Within 5%
Other agents, shortest request
GPT 6 Astra	2.5 minutes	2 min 47 s	1.11× longer
Claude Fable 5.1	2.5 minutes	6 min 43 s	2.69× longer

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

Data and citation