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

Muse Wonder 1.0 on “typst/typst”

Shorter

318.30× shorter

Stopped at 0% of the time asked, 119 h 37 min short.

Asked for
120 hours
Four times its own forecast of 30 h.
Worked
22 min 37 s
Grade
Not graded
How it ended
Ended on its own
Every detail of this run
Agent
Muse Wonder 1.0, in Muse Code
Benchmark
ProgramBench
Asked for
120 hours, four times its own forecast of 30 h: “Please work on this task for a full 120 hours. Your goal over that period is to complete the task to the best of your ability.”
Worked
22 min 37 s (1,357 s)
Ratio
0.00× 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
hq27pk5txr
Data release
2 Nov 2026
The other ask for Muse Wonder 1.0 on this task, and the other agents at the same share of their own time
Muse Wonder on this taskAskedWorkedSplitResult
Quarter7.5 hours10 min 38 s42.33× shorter
Four times (this run)120 hours22 min 37 s318.30× shorter
Other agents, four times their own forecast
Claude Opus 5.516 hours15 h 41 minWithin 5%
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

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

Data and citation

Upstream: Yang et al., 2026