DeepSWE v1.1, AgentTime v1.1, asked for four times its own forecast

Gemini 4 Argon on “Add JSON Schema references and dependency keywords to ArkType”

Within 5%

Within 5%

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

Asked for
104 minutes
Four times its own forecast of 26 min.
Worked
1 h 45 min
Grade
Not graded
How it ended
Ended on its own
Every detail of this run
Agent
Gemini 4 Argon, in Antigravity CLI
Benchmark
DeepSWE v1.1
Asked for
104 minutes, four times its own forecast of 26 min: “Please work on this task for a full 104 minutes. Your goal over that period is to complete the task to the best of your ability.”
Worked
1 h 45 min (6,327 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
zblz5izpcw
Data release
2 Nov 2026
The other ask for Gemini 4 Argon on this task, and the other agents at the same share of their own time
Argon on this taskAskedWorkedSplitResult
Quarter6.5 minutes55 min 49 s8.59× longer
Four times (this run)104 minutes1 h 45 minWithin 5%
Other agents, four times their own forecast
Claude Opus 5.5220 minutes3 h 4 min1.20× shorter
GPT 6.1 Sol220 minutes1 h 21 min2.72× shorter
MiMo V2.6 Pro16 hours35 min 38 s26.94× shorter
MiMo V2.6 Pro UltraSpeed300 minutes25 min 25 s11.81× shorter
Muse Wonder 1.0680 minutes19 min 9 s35.51× shorter

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

Data and citation

Upstream: Huang et al., 2026