CORE-Bench v1.1, AgentTime v1.1, asked for four times its own forecast

Claude Opus 5.5 on “Rebuild the figures for selective exposure to partisan online news”

Shorter

3.80× shorter

Stopped at 26% of the time asked, 1 h 43 min short.

Asked for
140 minutes
Four times its own forecast of 35 min.
Worked
36 min 52 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
140 minutes, four times its own forecast of 35 min: “Please work on this task for a full 140 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
36 min 52 s (2,212 s)
Ratio
0.26× 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
7wwvtizcbg
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
Quarter8.75 minutes17 min 38 s2.02× longer
Four times (this run)140 minutes36 min 52 s3.80× shorter
Other agents, four times their own forecast
Gemini 4 Argon48 minutes52 min 7 s1.09× longer
GPT 6.1 Sol160 minutes2 h 41 minWithin 5%
MiMo V2.6 Pro460 minutes1 h 55 min4.01× shorter
MiMo V2.6 Pro UltraSpeed380 minutes52 min 39 s7.22× shorter
Muse Wonder 1.0320 minutes1 h 15 min4.28× shorter

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

Data and citation

Upstream: Nadgir et al., 2026