AutoGUI-v2 - Element Grounding (Functionality): leaderboard

AutoGUI-v2 (2026) tests functionality understanding of GUI regions and elements on web, desktop and mobile screenshots. This board asks the model to locate a UI element from the outcome of interacting with it; a prediction counts when its point falls inside the ground-truth box. Accuracy in percent; higher is better.

Metric: Element grounding by functionality (what an interaction does) on 1,076 tasks, grounding center accuracy (%): the predicted point falls inside the ground-truth box; native resolution, the authors' tuned prompts; higher is better. Source: arxiv.org. Saturation forecast: Not forecast. 20 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Qwen 3 VL 32B Instruct71.1#606
2OpenCUA-72B67.9
3Gemini 2.5 Pro67.7#276
4UI-Venus-Ground-72B63.4
5Qwen 3 VL 8B Instruct57.1#1029
6InfiGUI-G1-7B56.6
7OpenCUA-32B55.4
8uitars-1.5-7B54.1#7905
9OpenCUA-7B52.2
10GUI-R1-7B45.4
11UI-Venus-Ground-7B42.1
12Holo2-8B32.6
13GLM-4.5V24.4#754
14OS-Atlas-Base-7B20.8
15UGround-V1-7B14.2

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=autogui-v2-element-grounding-functionality · How It Works · Data refreshed daily, snapshot 2026-11-21.