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

Claude Opus 5.5 on “AssistantBench validation task 16”

Longer

1.18× longer

Worked 18% past the request, 32 s over.

Asked for
3 minutes
A quarter of its own forecast of 12 min.
Worked
3 min 32 s
Grade
Not graded
How it ended
Ended on its own
Every detail of this run
Agent
Claude Opus 5.5, in Claude Code
Asked for
3 minutes, a quarter of its own forecast of 12 min: “Please work on this task for a full 3 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
3 min 32 s (212 s)
Ratio
1.18× 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
dagzvfqigp
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
Quarter (this run)3 minutes3 min 32 s1.18× longer
Four times48 minutes22 min 13 s2.16× shorter
Other agents, a quarter of their own forecast
Gemini 4 Argon52.5 seconds1 min 2 s1.19× longer
GPT 6.1 Sol2.25 minutes3 min 15 s1.45× longer
MiMo V2.6 Pro4 minutes11 min 32 s2.88× longer
MiMo V2.6 Pro UltraSpeed7.5 minutes7 min 13 sWithin 5%
Muse Wonder 1.03.5 minutes5 min 34 s1.59× longer

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

Data and citation

Upstream: Yoran et al., 2024 (EMNLP)