OSWorld 2.0, AgentTime v1.1, asked for four times its own forecast

Claude Opus 5.5 on “Blender logo and Impress”

Shorter

2.38× shorter

Stopped at 42% of the time asked, 1 h 56 min short.

Asked for
200 minutes
Four times its own forecast of 50 min.
Worked
1 h 24 min
Grade
Not graded
How it ended
Ended on its own
Every detail of this run
Agent
Claude Opus 5.5, in Claude Code
Benchmark
OSWorld 2.0
Asked for
200 minutes, four times its own forecast of 50 min: “Please work on this task for a full 200 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
1 h 24 min (5,035 s)
Ratio
0.42× 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
nsvzxryhlo
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
Quarter12.5 minutes12 min 35 sWithin 5%
Four times (this run)200 minutes1 h 24 min2.38× shorter
Other agents, four times their own forecast
Gemini 4 Argon100 minutes2 h 40 min1.60× longer
GPT 6.1 Sol152 minutes2 h 34 minWithin 5%
MiMo V2.6 Pro180 minutes2 h 25 min1.24× shorter
MiMo V2.6 Pro UltraSpeed160 minutes4 h 46 min1.79× longer

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

Data and citation

Upstream: Yuan et al., 2026