ParaGUIBench - Web Search: leaderboard
Metric: Success rate (%; 65 web-search tasks; single serial GUI agent (one perception-decision-action loop, no planner) on a Docker desktop with 4 GiB RAM and 2 vCPUs, each model under its official visual-history setting; verdict read from the final environment state by task-specific checks). Source: arxiv.org. Saturation forecast: Around July 2027. 4 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4.6 | 32.3 |
| 2 | Seed 1.8 | 24.6 |
| 3 | Qwen 3 VL 235B A22B Instruct | 12.3 |
Interactive version: theaggregate.ai/benchmark?slug=paraguibench-web-search · How It Works · Data refreshed daily, snapshot 2026-09-29.