DeepSWE v1.1, AgentTime v1.1, asked for a quarter of its own forecast

Claude Opus 5.5 on “Add session bundle recording and replay to IPython”

Shorter

2.22× shorter

Stopped at 45% of the time asked, 4 min 49 s short.

Asked for
8.75 minutes
A quarter of its own forecast of 35 min.
Worked
3 min 56 s
Grade
Not graded
How it ended
Ended on its own
Every detail of this run
Agent
Claude Opus 5.5, in Claude Code
Benchmark
DeepSWE v1.1
Asked for
8.75 minutes, a quarter of its own forecast of 35 min: “Please work on this task for a full 8.75 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
3 min 56 s (236 s)
Ratio
0.45× 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
ydjrbtytvh
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)8.75 minutes3 min 56 s2.22× shorter
Four times140 minutes1 h 7 min2.10× shorter
Other agents, a quarter of their own forecast
Gemini 4 Argon3.75 minutes14 min 17 s3.81× longer
GPT 6.1 Sol10 minutes10 min 16 sWithin 5%
MiMo V2.6 Pro1 hour19 min 4 s3.15× shorter
MiMo V2.6 Pro UltraSpeed46.25 minutes6 min 29 s7.13× shorter
Muse Wonder 1.027.5 minutes11 min 11 s2.46× shorter

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

Data and citation

Upstream: Huang et al., 2026