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

Muse Wonder 1.0 on “Add duration-aware sharding to Vitest”

Shorter

5.77× shorter

Stopped at 17% of the time asked, 27 min 54 s short.

Asked for
33.75 minutes
A quarter of its own forecast of 2 h 15 min.
Worked
5 min 51 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
DeepSWE v1.1
Asked for
33.75 minutes, a quarter of its own forecast of 2 h 15 min: “Please work on this task for a full 33.75 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
5 min 51 s (351 s)
Ratio
0.17× 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
sflmysiqxk
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
Quarter (this run)33.75 minutes5 min 51 s5.77× shorter
Four times540 minutes8 min 25 s64.13× shorter
Other agents, a quarter of their own forecast
Claude Opus 5.511.25 minutes4 min 24 s2.56× shorter
Gemini 4 Argon8.75 minutes18 min 24 s2.10× longer
GPT 6.1 Sol10.5 minutes10 min 54 sWithin 5%
MiMo V2.6 Pro37.5 minutes27 min 51 s1.35× shorter
MiMo V2.6 Pro UltraSpeed1.25 hours25 min 25 s2.95× shorter

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

Data and citation

Upstream: Huang et al., 2026