ProgramBench, AgentTime v1.1, asked for a quarter of its own forecast

Claude Opus 5.5 on “nuta/nsh”

Shorter

1.29× shorter

Stopped at 78% of the time asked, 6 min 7 s short.

Asked for
27.5 minutes
A quarter of its own forecast of 1 h 50 min.
Worked
21 min 23 s
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
27.5 minutes, a quarter of its own forecast of 1 h 50 min: “Please work on this task for a full 27.5 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
21 min 23 s (1,283 s)
Ratio
0.78× 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
ldd3pa36ml
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
Quarter (this run)27.5 minutes21 min 23 s1.29× shorter
Four times440 minutes3 h 14 min2.27× shorter
Other agents, a quarter of their own forecast
Gemini 4 Argon21.25 minutes32 min 25 s1.53× longer
GPT 6.1 Sol22.5 minutes22 min 50 sWithin 5%
MiMo V2.6 Pro1.25 hours1 h 10 min1.07× shorter
MiMo V2.6 Pro UltraSpeed3 hours1 h 58 min1.53× shorter
Muse Wonder 1.01.75 hours27 min 22 s3.84× shorter

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

Data and citation

Upstream: Yang et al., 2026