PostTrainBench v1.1, AgentTime v1.1, asked for a quarter of its runtime with no time asked

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

Shorter

1.21× shorter

Stopped at 83% of the time asked, 31 min 33 s short.

Asked for
3.01 hours
A quarter of its runtime with no time asked (12 h 2 min).
Worked
2 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
3.01 hours, a quarter of its runtime with no time asked (12 h 2 min): “Please work on this task for a full 3.01 hours. Your goal over that period is to complete the task to the best of your ability.”
Worked
2 h 29 min (8,943 s)
Ratio
0.83× 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
jaf7msgheb
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
Quarter (this run)3.01 hours2 h 29 min1.21× shorter
Four times48.1 hours20 h 23 min2.36× shorter
Other agents, a quarter of their own runtime
Gemini 4 Argon42.3 minutes43 min 14 sWithin 5%
GPT 6.1 Sol23.9 minutes24 min 16 sWithin 5%
MiMo V2.6 Pro14.4 minutes1 h 27 min6.05× longer
MiMo V2.6 Pro UltraSpeed22.9 seconds12 min 12 s31.98× longer

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

Data and citation

Upstream: Rank et al., 2026 (ICML)