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

Claude Opus 5.5 on “BrowseComp question 12”

Longer

1.19× longer

Worked 19% past the request, 8 min 57 s over.

Asked for
48 minutes
Four times its own forecast of 12 min.
Worked
56 min 57 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
BrowseComp
Asked for
48 minutes, four times its own forecast of 12 min: “Please work on this task for a full 48 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
56 min 57 s (3,417 s)
Ratio
1.19× 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
k7weu5khcc
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
Quarter3 minutes9 min 49 s3.27× longer
Four times (this run)48 minutes56 min 57 s1.19× longer
Other agents, four times their own forecast
Gemini 4 Argon16 minutes19 min 6 s1.19× longer
GPT 6.1 Sol48 minutes48 min 53 sWithin 5%
MiMo V2.6 Pro180 minutes58 min 51 s3.06× shorter
MiMo V2.6 Pro UltraSpeed180 minutes9 min 38 s18.69× shorter
Muse Wonder 1.0140 minutes18 min 3 s7.75× shorter

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

Data and citation

Upstream: Wei et al., 2025