ParaGUIBench: leaderboard
ParaGUIBench (2026) is a benchmark of 233 long-horizon desktop and web GUI tasks whose sub-tasks can often run in parallel; this board scores single serial GUI agents with one perception-decision-action loop. Success rate in percent over all 233 tasks; higher is better.
Metric: Success rate (%; all 233 long-horizon desktop and web tasks with parallelizable sub-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 February 2027. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 33.5 | #161 |
| 2 | Seed 1.8 | 27.9 | #242 |
| 3 | Qwen 3 VL 235B A22B Instruct | 11.2 | #522 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=paraguibench · How It Works · Data refreshed daily, snapshot 2026-11-13.