CAP (Cross-Site Browser Tasks) - Complex Actions: leaderboard
The complex-action score of CAP (Macau University of Science and Technology, FellouAI and others, 2026): the mean score of execution checkpoints that need non-trivial UI operations such as map dragging, sliders and video controls, over a fixed 198-task set of cross-site browser tasks. Percent, higher is better.
Metric: Action score (%; mean leaf score over the execution checkpoints for complex UI operations such as map dragging, sliders and video control; a fixed 198-task evaluation set of CAP cross-site browser tasks on live websites (the benchmark holds 420 tasks over 108 websites in 24 domains); each task becomes a hierarchical rubric tree of execution, perception and format checkpoints, scored by a GPT-4o judge agent with URL-grounded verification; Browser-Use agent framework with default configuration and at most 50 reasoning-action steps). Source: arxiv.org. Saturation forecast: Around June 2027. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V4 Flash | 33.7 | #258 |
| 2 | Claude Sonnet 4.5 | 32 | #250 |
| 3 | GPT-5 | 29 | #167 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=cap-cross-site-browser-tasks-complex-actions · How It Works · Data refreshed daily, snapshot 2026-11-13.