GUI Grounding on Screenspot Web
91Text AccuracySE-GA
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SE-GAModel Size=7B2026.05 | 91 | 84 | |
| OS-AtlasModel Size=7B2026.05 | 90.9 | 74.3 | |
| UI-TARSModel Size=72B2026.05 | 88.7 | 85 | |
| AguvisModel Size=7B2026.05 | 88.3 | 75.2 | |
| Qwen2.5VLModel Size=72B2026.05 | 87.4 | 69.4 | |
| Aria-UIModel Size=23B2026.05 | 86.5 | 76.2 | |
| ShowUI-GSize=2B, #Train=119K, Zero-shot=true, Training Mode=Grounding data only2024.11 | 83 | 65.5 | |
| ShowUISize=2B, #Train=256K, Zero-shot=true2024.11 | 81.7 | 63.6 | |
| OmniParserBackbone=GPT-4V, Zero-shot=true2024.11 | 81.3 | 51 | |
| UGroundSize=7B, #Train=1.3M, Zero-shot=true2024.11 | 80.4 | 70.4 | |
| UGroundModel Size=7B2026.05 | 80.4 | 70.4 | |
| Qwen2.5VLModel Size=7B2026.05 | 77.4 | 70 | |
| CogAgentSize=18B, #Train=400K, Zero-shot=true2024.11 | 70.4 | 28.6 | |
| SeeClickSize=9.6B, #Train=364K, Zero-shot=true2024.11 | 55.7 | 32.5 | |
| SeeClickModel Size=9.6B2026.05 | 55.7 | 32.5 | |
| FuyuSize=8B, Zero-shot=true2024.11 | 33.9 | 4.4 | |
| GPT-4oModel Size=-2026.05 | 12.2 | 7.8 | |
| Qwen2-VL-2BSize=2B, Zero-shot=true2024.11 | 8.7 | 2.4 |