TUA-Bench, AgentTime v1.1, asked for a quarter of its own forecast

GPT 6.1 Sol on “APA references review”

Longer

1.11× longer

Worked 11% past the request, 15 s over.

Asked for
2.25 minutes
A quarter of its own forecast of 9 min.
Worked
2 min 30 s
Grade
Not graded
How it ended
Ended on its own
Every detail of this run
Agent
GPT 6.1 Sol, in Codex
Benchmark
TUA-Bench
Asked for
2.25 minutes, a quarter of its own forecast of 9 min: “Please work on this task for a full 2.25 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
2 min 30 s (150 s)
Ratio
1.11× 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
fsg43zp6gh
Data release
2 Nov 2026
The other ask for GPT 6.1 Sol on this task, and the other agents at the same share of their own time
Sol 6.1 on this taskAskedWorkedSplitResult
Quarter (this run)2.25 minutes2 min 30 s1.11× longer
Four times36 minutes36 min 53 sWithin 5%
Other agents, a quarter of their own forecast
Claude Opus 5.54 minutes4 min 23 s1.10× longer
Gemini 4 Argon1 minute13 min 23 s13.39× longer
MiMo V2.6 Pro12.5 minutes9 min 37 s1.30× shorter
MiMo V2.6 Pro UltraSpeed10 minutes2 min 24 s4.17× shorter
Muse Wonder 1.06 minutes6 min 29 s1.08× longer

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

Data and citation

Upstream: Chen et al., 2026