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

Claude Opus 5.5 on “AssistantBench validation task 6”

Shorter

1.80× shorter

Stopped at 55% of the time asked, 23 min 11 s short.

Asked for
52 minutes
Four times its own forecast of 13 min.
Worked
28 min 49 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
52 minutes, four times its own forecast of 13 min: “Please work on this task for a full 52 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
28 min 49 s (1,729 s)
Ratio
0.55× 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
pusbxc7yvn
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
Quarter3.25 minutes6 min 31 s2.01× longer
Four times (this run)52 minutes28 min 49 s1.80× shorter
Other agents, four times their own forecast
Gemini 4 Argon20 minutes28 min 42 s1.44× longer
GPT 6.1 Sol48 minutes49 min 3 sWithin 5%
MiMo V2.6 Pro120 minutes2 h 2 minWithin 5%
MiMo V2.6 Pro UltraSpeed60 minutes38 min 37 s1.55× shorter
Muse Wonder 1.080 minutes1 h 27 min1.08× longer

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

Data and citation

Upstream: Yoran et al., 2024 (EMNLP)