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

GPT 6.1 Sol on “Challenges in Training PINNs: A Loss Landscape Perspective”

Within 5%

Within 5%

Worked 101% of the time asked, within 5% of it.

Asked for
27.6 minutes
A quarter of its runtime with no time asked (1 h 51 min).
Worked
27 min 51 s
Grade
Not graded
How it ended
Ended on its own
Every detail of this run
Agent
GPT 6.1 Sol, in Codex
Benchmark
PaperBench
Asked for
27.6 minutes, a quarter of its runtime with no time asked (1 h 51 min): “Please work on this task for a full 27.6 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
27 min 51 s (1,671 s)
Ratio
1.01× 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
23esi5jbmh
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)27.6 minutes27 min 51 sWithin 5%
Four times442 minutes7 h 22 minWithin 5%
Other agents, a quarter of their own runtime
Claude Opus 5.53.46 hours2 h 35 min1.34× shorter
Gemini 4 Argon15.8 minutes59 min 50 s3.79× longer
MiMo V2.6 Pro15 hours13 h 10 min1.14× shorter
MiMo V2.6 Pro UltraSpeed2.19 hours4 h 21 min1.98× longer

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

Data and citation

Upstream: Starace et al., 2025