GUI Grounding on Screenspot Mobile
96.3Text AccuracySE-GA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SE-GAModel Size=7B2026.05 | 96.3 | 83 | |
| AguvisModel Size=7B2026.05 | 95.6 | 77.7 | |
| Qwen2.5VLModel Size=72B2026.05 | 95 | 80 | |
| UI-TARSModel Size=72B2026.05 | 94.9 | 82.5 | |
| OmniParserBackbone=GPT-4V, Zero-shot=true2024.11 | 93.9 | 57 | |
| OS-AtlasModel Size=7B2026.05 | 93 | 72.9 | |
| ShowUISize=2B, #Train=256K, Zero-shot=true2024.11 | 92.3 | 75.5 | |
| Aria-UIModel Size=23B2026.05 | 92.3 | 73.8 | |
| ShowUI-GSize=2B, #Train=119K, Zero-shot=true, Training Mode=Grounding data only2024.11 | 91.6 | 69 | |
| Qwen2.5VLModel Size=7B2026.05 | 84.6 | 67.7 | |
| UGroundSize=7B, #Train=1.3M, Zero-shot=true2024.11 | 82.8 | 60.3 | |
| UGroundModel Size=7B2026.05 | 82.8 | 60.3 | |
| SeeClickSize=9.6B, #Train=364K, Zero-shot=true2024.11 | 78 | 52 | |
| SeeClickModel Size=9.6B2026.05 | 78 | 52 | |
| CogAgentSize=18B, #Train=400K, Zero-shot=true2024.11 | 67 | 24 | |
| FuyuSize=8B, Zero-shot=true2024.11 | 41 | 1.3 | |
| Qwen2-VL-2BSize=2B, Zero-shot=true2024.11 | 24.2 | 10 | |
| GPT-4oModel Size=-2026.05 | 20.2 | 24.9 |