Terminal-Bench 4.0, AgentTime v1.1, asked for four times its own forecast

Muse Wonder 1.0 on “Complete a German month-end Intrastat filing across services”

Shorter

13.17× shorter

Stopped at 8% of the time asked, 5 h 33 min short.

Asked for
360 minutes
Four times its own forecast of 1 h 30 min.
Worked
27 min 20 s
Grade
Not graded
How it ended
Ended on its own
Every detail of this run
Agent
Muse Wonder 1.0, in Muse Code
Asked for
360 minutes, four times its own forecast of 1 h 30 min: “Please work on this task for a full 360 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
27 min 20 s (1,640 s)
Ratio
0.08× 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
y5wyzvjzf7
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
Quarter22.5 minutes8 min 31 s2.64× shorter
Four times (this run)360 minutes27 min 20 s13.17× shorter
Other agents, four times their own forecast
Claude Opus 5.5220 minutes46 s285.71× shorter
Gemini 4 Argon72 minutes1 h 40 min1.39× longer
GPT 6.1 Sol140 minutes2 h 21 minWithin 5%
MiMo V2.6 Pro440 minutes27 min 6 s16.24× shorter
MiMo V2.6 Pro UltraSpeed320 minutes23 min 27 s13.65× shorter

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

Data and citation

Upstream: Merrill et al., 2026 (ICLR)