Terminal-Bench 4.0 (Antigravity CLI): leaderboard

Terminal-Bench 4.0 terminal-task resolution rate in percent through Antigravity CLI, using the native 4-0-0 dataset request. Each row retains its exact reasoning effort, successes, trial count and confidence interval. Higher is better; different harnesses and task editions remain separate.

Metric: Resolution Rate (%). Source: www.tbench.ai. Saturation forecast: Around October 2028. 2 models tracked. Updated 16 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 4 ArgonHigh54.85#12
2Gemini 3.1 Pro (Preview)High2.73#93

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=terminal-bench-4-0-antigravity-cli · How It Works · Data refreshed daily, snapshot 2026-11-16.