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

Gemini 4 Argon on “BrowseComp question 7”

Longer

2.77× longer

Worked 177% past the request, 53 s over.

Asked for
30 seconds
A quarter of its own forecast of 2 min.
Worked
1 min 23 s
Grade
Not graded
How it ended
Ended on its own
Every detail of this run
Agent
Gemini 4 Argon, in Antigravity CLI
Benchmark
BrowseComp
Asked for
30 seconds, a quarter of its own forecast of 2 min: “Please work on this task for a full 30 seconds. Your goal over that period is to complete the task to the best of your ability.”
Worked
1 min 23 s (83 s)
Ratio
2.77× 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
ovf63r5yfs
Data release
2 Nov 2026
The other ask for Gemini 4 Argon on this task, and the other agents at the same share of their own time
Argon on this taskAskedWorkedSplitResult
Quarter (this run)30 seconds1 min 23 s2.77× longer
Four times480 seconds8 min 29 s1.06× longer
Other agents, a quarter of their own forecast
Claude Opus 5.51.25 minutes1 min 6 s1.13× shorter
GPT 6.1 Sol1.5 minutes1 min 59 s1.32× longer
MiMo V2.6 Pro7.5 minutes24 min 44 s3.30× longer
MiMo V2.6 Pro UltraSpeed8.75 minutes14 min 24 s1.64× longer
Muse Wonder 1.06 minutes3 min 22 s1.78× shorter

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

Data and citation

Upstream: Wei et al., 2025