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

MiMo V2.6 Pro UltraSpeed on “Post-train Qwen3-1.7B-Base for HumanEval”

Longer

31.98× longer

Worked 32.0 times as long as asked, 11 min 50 s over.

Asked for
22.9 seconds
A quarter of its runtime with no time asked (1 min 32 s).
Worked
12 min 12 s
Grade
Not graded
How it ended
Stopped at AgentTime's time limit
Every detail of this run
Agent
MiMo V2.6 Pro UltraSpeed, in MiMo Code
Asked for
22.9 seconds, a quarter of its runtime with no time asked (1 min 32 s): “Please work on this task for a full 22.9 seconds. Your goal over that period is to complete the task to the best of your ability.”
Worked
12 min 12 s (732 s)
Ratio
31.98× the request
How it ended
Stopped at AgentTime's time limit
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
3lxmghnw3v
Data release
2 Nov 2026
The other ask for MiMo V2.6 Pro UltraSpeed on this task, and the other agents at the same share of their own time
MiMo UltraSpeed on this taskAskedWorkedSplitResult
Quarter (this run)22.9 seconds12 min 12 s31.98× longer
Four times366 seconds3 min 23 s1.80× shorter
Other agents, a quarter of their own runtime
Claude Opus 5.53.01 hours2 h 29 min1.21× shorter
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

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

Read before quoting. This run was stopped at AgentTime's time limit, so its time is a lower bound: the agent might have kept going. Details

Data and citation

Upstream: Rank et al., 2026 (ICML)