PostTrainBench v1.1, AgentTime v1.1, asked for four times its runtime with no time asked

Claude Opus 5.5 on “Post-train Qwen3-1.7B-Base for HumanEval”

Shorter

2.36× shorter

Stopped at 42% of the time asked, 27 h 43 min short.

Asked for
48.1 hours
Four times its runtime with no time asked (12 h 2 min).
Worked
20 h 23 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
48.1 hours, four times its runtime with no time asked (12 h 2 min): “Please work on this task for a full 48.1 hours. Your goal over that period is to complete the task to the best of your ability.”
Worked
20 h 23 min (73,378 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
f5fjgdyl6f
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
Quarter3.01 hours2 h 29 min1.21× shorter
Four times (this run)48.1 hours20 h 23 min2.36× shorter
Other agents, four times their own runtime
Gemini 4 Argon677 minutes4 h 6 min2.75× shorter
GPT 6.1 Sol383 minutes6 h 24 minWithin 5%
MiMo V2.6 Pro230 minutes2 h 36 min1.47× shorter
MiMo V2.6 Pro UltraSpeed366 seconds3 min 23 s1.80× shorter

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

Data and citation

Upstream: Rank et al., 2026 (ICML)