Terminal-Bench 4.0, AgentTime v1.1, asked for four times its own forecast

Claude Opus 5.5 on “Repair a React lead form and its auditable CRM submission pipeline”

Shorter

2.02× shorter

Stopped at 49% of the time asked, 1 h 31 min short.

Asked for
180 minutes
Four times its own forecast of 45 min.
Worked
1 h 29 min
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
180 minutes, four times its own forecast of 45 min: “Please work on this task for a full 180 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
1 h 29 min (5,342 s)
Ratio
0.49× 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
o2jut2gbnf
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
Quarter11.25 minutes12 min 23 s1.10× longer
Four times (this run)180 minutes1 h 29 min2.02× shorter
Other agents, four times their own forecast
Gemini 4 Argon72 minutes28 min 50 s2.50× shorter
GPT 6.1 Sol120 minutes2 h 1 minWithin 5%
MiMo V2.6 Pro840 minutes37 min 21 s22.49× shorter
MiMo V2.6 Pro UltraSpeed16 hours16 min 3 s59.81× shorter
Muse Wonder 1.0340 minutes9 min 59 s34.03× shorter

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

Data and citation

Upstream: Merrill et al., 2026 (ICLR)