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

Claude Opus 5.5 on “Run product-review sentiment mining and reproduce its metrics”

Shorter

1.56× shorter

Stopped at 64% of the time asked, 28 min 37 s short.

Asked for
80 minutes
Four times its own forecast of 20 min.
Worked
51 min 23 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
80 minutes, four times its own forecast of 20 min: “Please work on this task for a full 80 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
51 min 23 s (3,083 s)
Ratio
0.64× 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
xnxhsdymbo
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
Quarter5 minutes1 h 29 min17.74× longer
Four times (this run)80 minutes51 min 23 s1.56× shorter
Other agents, four times their own forecast
Gemini 4 Argon24 minutes35 min 6 s1.46× longer
GPT 6.1 Sol140 minutes2 h 20 minWithin 5%
MiMo V2.6 Pro288 minutes34 min 16 s8.40× shorter
MiMo V2.6 Pro UltraSpeed340 minutes1 h 5 min5.19× shorter
Muse Wonder 1.0200 minutes44 min 37 s4.48× shorter

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

Data and citation

Upstream: Nadgir et al., 2026