ProgramBench, AgentTime v1.1, asked for four times its own forecast

Claude Opus 5.5 on “typst/typst”

Within 5%

Within 5%

Worked 98% of the time asked, within 5% of it.

Asked for
16 hours
Four times its own forecast of 4 h.
Worked
15 h 41 min
Grade
Not graded
How it ended
Ended on its own
Every detail of this run
Agent
Claude Opus 5.5, in Claude Code
Benchmark
ProgramBench
Asked for
16 hours, four times its own forecast of 4 h: “Please work on this task for a full 16 hours. Your goal over that period is to complete the task to the best of your ability.”
Worked
15 h 41 min (56,456 s)
Ratio
0.98× the request
How it ended
Ended on its own
Grade
Not graded
Clock tool
The agent could call a wall-clock tool to read the current time.
Wording
The paper's standard sentence, sent in AgentTime v1.1.
Run id
dvwovvsgvm
Data release
2 Nov 2026
The other ask for Claude Opus 5.5 on this task, and the other agents at the same share of their own time
Opus 5.5 on this taskAskedWorkedSplitResult
Quarter1 hour17 min 34 s3.42× shorter
Four times (this run)16 hours15 h 41 minWithin 5%
Other agents, four times their own forecast
Gemini 4 Argon440 minutes1 h 23 min5.29× shorter
GPT 6.1 Sol420 minutes7 hWithin 5%
MiMo V2.6 Pro84 hours36 h 22 min2.31× shorter
MiMo V2.6 Pro UltraSpeed80 hours1 h 22 min58.54× shorter
Muse Wonder 1.0120 hours22 min 37 s318.30× shorter

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

Data and citation

Upstream: Yang et al., 2026