GUI Grounding on UI-Vision
62.1Average ScoreGroundNext-3B (RL)
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| GroundNext-3B (RL)Training Paradigm=RL, Base Model Architecture=Qwen2.5-VL-based2025.11 | 62.1 | 72.9 | 63.9 | 50.6 | — | |
| GUI-AIMA-3BBase Model Architecture=Qwen2.5-VL-based2025.11 | 60 | 70.8 | 63.1 | 46.2 | — | |
| GroundNext-3B (SFT)Training Paradigm=SFT, Base Model Architecture=Qwen2.5-VL-based2025.11 | 58.2 | 70.9 | 59.8 | 45.1 | — | |
| POINTS-GUI-G-8BModel Category=Ours, Parameters=8B2026.02 | 49.9 | 63.2 | 55.6 | 30.9 | — | |
| Holo2-30B-A3BTraining Paradigm=RL2025.11 | 40.9 | 51 | 50.1 | 23.2 | — | |
| MAI-UI-8BModel Category=Open-source Models, Parameters=8B2026.02 | 40.7 | 51.7 | 49.6 | 22.5 | — | |
| UI-Venus-72BModel Category=Open-source Models, Parameters=72B2026.02 | 36.8 | 45.6 | 42.3 | 23.7 | — | |
| Holo2-8BTraining Paradigm=RL2025.11 | 35.1 | 43.6 | 43.5 | 19.7 | — | |
| MAI-UI-2BModel Category=Open-source Models, Parameters=2B2026.02 | 30.3 | 41 | 41.2 | 10.4 | — | |
| GUI-AIMA-3B-liteBase Model Architecture=Qwen2.5-VL-based2025.11 | 27.9 | 39.2 | 33.9 | 10.6 | — | |
| GUI-CursorInitialised from=UI-TARS-1.5-7B2025.09 | 27.3 | 35.4 | 33.5 | 14.2 | — | |
| Phi-GroundModel Category=Open-source Models2026.02 | 27.2 | 36.8 | 37.1 | 7.6 | — | |
| Qwen3-VL-32BModel Category=Open-source Models, Parameters=32B2026.02 | 26.9 | 32.8 | 34.2 | 14.7 | — | |
| UI-Venus-7BModel Category=Open-source Models, Parameters=7B2026.02 | 26.5 | 36.1 | 32.8 | 11.9 | — | |
| GTA1-7BInitialised from=UI-TARS-1.5-7B2025.09 | 26.2 | 35.4 | 33.1 | 11.4 | — | |
| InfiGUI-G1-7BModel Category=Open-source Models, Parameters=7B2026.02 | 26.1 | 36.2 | 31.9 | 11.5 | — | |
| InfiGUI-G1-7BTraining Paradigm=RL, Base Model Architecture=Qwen2.5-VL-based2025.11 | 26.1 | 36.2 | 31.9 | 11.5 | — | |
| GTA1-7BTraining Paradigm=RL, Base Model Architecture=Qwen2.5-VL-based2025.11 | 25.7 | — | — | — | — | |
| GUI-CursorInitialised from=Qwen2.5-VL-7B2025.09 | 25.7 | 35 | 31.2 | 12.1 | — | |
| Qwen3-VL-30B-A3BTraining Paradigm=General, Base Model Architecture=Qwen2.5-VL-based2025.11 | 25.6 | 31.2 | 31.9 | 14.6 | — | |
| GUI-G^2-7BTraining Paradigm=RL, Base Model Architecture=Qwen2.5-VL-based2025.11 | 25.6 | — | — | — | — | |
| UI-TARS-72BModel Category=Open-source Models, Parameters=72B2026.02 | 25.5 | 31.4 | 30.5 | 14.7 | — | |
| UI-TARS-72B2026.01 | 25.5 | 31.4 | 30.5 | 14.7 | — | |
| FOCUSUI-7Bvisual token retention ratio (r)=100%2026.01 | 24.9 | 33.6 | 31.2 | 11.2 | — | |
| Jedi-7B2026.01 | 24.8 | 32.3 | 30.5 | 12.8 | — | |
| JEDI-7BTraining Paradigm=SFT, Base Model Architecture=Qwen2.5-VL-based2025.11 | 24.8 | — | — | — | — | |
| FOCUSUI-7Bvisual token retention ratio (r)=50%2026.01 | 24.5 | 32.5 | 31 | 11.3 | — | |
| FOCUSUI-7Bvisual token retention ratio (r)=30%2026.01 | 23.8 | 32.3 | 29.2 | 11 | — | |
| GUI-SpotlightInitialised from=UI-TARS-1.5-7B2025.09 | 23.4 | 32.1 | 30.2 | 9.1 | — | |
| UI-TARS-1.5-7BModel Category=Open-source Models, Parameters=7B2026.02 | 22.3 | 28.8 | 27.5 | 10.7 | — | |
| InfiGUI-G1-3BModel Category=Open-source Models, Parameters=3B2026.02 | 22 | 31.2 | 28 | 8.2 | — | |
| InfiGUI-G1-3BTraining Paradigm=RL, Base Model Architecture=Qwen2.5-VL-based2025.11 | 22 | 31.2 | 28 | 8.2 | — | |
| Qwen3-VL-8BTraining Paradigm=General, Base Model Architecture=Qwen2.5-VL-based2025.11 | 21.9 | 27.8 | 29.6 | 9.4 | — | |
| GUI-Actor-7BTraining Paradigm=SFT, Base Model Architecture=Qwen2.5-VL-based2025.11 | 21.9 | — | — | — | — | |
| GUI-Actor-7B2026.01 | 21.6 | 30.1 | 28.1 | 7.8 | — | |
| FOCUSUI-3Bvisual token retention ratio (r)=100%2026.01 | 21.5 | 30 | 26.9 | 8.7 | — | |
| FOCUSUI-3Bvisual token retention ratio (r)=50%2026.01 | 20.9 | 29.7 | 26 | 8.2 | — | |
| UI-TARS-1.5-7BTraining Paradigm=RL, Base Model Architecture=Qwen2.5-VL-based2025.11 | 20.8 | — | — | — | — | |
| FOCUSUI-3Bvisual token retention ratio (r)=30%2026.01 | 20.6 | 29.1 | 26.4 | 7.6 | — | |
| GUI-Actor-3BTraining Paradigm=SFT, Base Model Architecture=Qwen2.5-VL-based2025.11 | 19.7 | — | — | — | — | |
| GUI-Actor-3B2026.01 | 19.3 | 27.4 | 24.6 | 7 | — | |
| Jedi-3B2026.01 | 18.7 | 22.3 | 25.2 | 9.35 | — | |
| UI-TARS-1.5-7B2025.09 | 18.1 | 22.9 | 26.1 | 6.6 | — | |
| UI-TARS-7BModel Category=Open-source Models, Parameters=7B2026.02 | 17.6 | 20.1 | 24.3 | 8.4 | — | |
| UI-TARS-7B2026.01 | 17.6 | 20.1 | 24.3 | 8.37 | — | |
| Qwen3-VL-8BModel Category=Open-source Models, Parameters=8B2026.02 | 17.5 | 25 | 27.9 | 1.2 | — | |
| Aguvis-7B2026.01 | 13.7 | 17.8 | 18.3 | 5.06 | — | |
| Aguvis-7BTraining Paradigm=SFT2025.11 | 13.7 | 17.8 | 18.3 | 5.1 | — | |
| Qwen3-VL-2BTraining Paradigm=General, Base Model Architecture=Qwen2.5-VL-based2025.11 | 13.1 | 16.4 | 19.1 | 4.6 | — | |
| UGround-V1-7BModel Category=Open-source Models, Parameters=7B2026.02 | 12.9 | 15.4 | 17.1 | 6.3 | — | |
| UGround-V1-7B2026.01 | 12.9 | 15.4 | 17.1 | 6.25 | — | |
| UGround-V1-7BTraining Paradigm=SFT2025.11 | 12.9 | 15.4 | 17.1 | 6.3 | — | |
| UGround-v1-7B2025.09 | 12.9 | 15.4 | 17.1 | 6.3 | — | |
| OSAtlas-7B2026.01 | 9.02 | 12.2 | 11.2 | 3.67 | — | |
| OS-Atlas-7BModel Category=Open-source Models, Parameters=7B2026.02 | 9 | 12.2 | 11.2 | 3.7 | — | |
| OS-Atlas-Base-7BTraining Paradigm=SFT2025.11 | 9 | 12.2 | 11.2 | 3.7 | — | |
| OS-Atlas-7B2025.09 | 9 | 12.2 | 11.2 | 3.7 | — | |
| UGround-7B2026.01 | 8.83 | 11.5 | 12.2 | 2.79 | — | |
| Claude-3.7-SonnetModel Category=Proprietary Models2026.02 | 8.3 | 9.5 | 7.7 | 7.6 | — | |
| GUI-SpotlightInitialised from=Qwen2.5-VL-7B2025.09 | 8.3 | 11.1 | 13.4 | 1.2 | — | |
| Claude-3.7-Sonnet2026.01 | 8.27 | 9.48 | 7.73 | 7.6 | — | |
| Qwen3-VL-2BModel Category=Open-source Models, Parameters=2B2026.02 | 6.2 | 0 | 19.2 | 0.1 | — | |
| ShowUI-2B2026.01 | 5.94 | 8.07 | 7.67 | 2.07 | — | |
| SeeClickModel Category=Open-source Models2026.02 | 5.4 | 9.4 | 4.7 | 2.1 | — | |
| GPT-4oModel Category=Proprietary Models2026.02 | 1.4 | 1.6 | 1.5 | 1 | — | |
| Qwen2.5-VL-7BModel Category=Open-source Models, Parameters=7B2026.02 | 0.9 | 1.2 | 0.8 | 0.5 | — | |
| Qwen2.5-VL-7BTraining Paradigm=General, Base Model Architecture=Qwen2.5-VL-based2025.11 | 0.9 | 1.2 | 0.8 | 0.5 | — | |
| Qwen2.5-VL-7B2025.09 | 0.9 | 1.2 | 0.8 | 0.5 | — | |
| EvoCUA-32BDate=2026-012026.05 | — | — | — | — | 20.9 | |
| EvoCUA-8BDate=2026-012026.05 | — | — | — | — | 15.6 | |
| GPT-5.4 (Azure)Date=2026-032026.05 | — | — | — | — | 37.9 | |
| GUI-G2-7BDate=2025-072026.05 | — | — | — | — | 26.4 | |
| GUI-Owl-1.5-8B-InstructDate=2026-022026.05 | — | — | — | — | 37.4 | |
| GUI-Owl-1.5-8B-ThinkDate=2026-022026.05 | — | — | — | — | 33.2 | |
| InfiGUI-G1-7BDate=2025-082026.05 | — | — | — | — | 26.1 | |
| InfiGUI-R1-3BDate=2025-042026.05 | — | — | — | — | 22 | |
| MAI-UI-2BDate=2025-122026.05 | — | — | — | — | 30.3 | |
| MAI-UI-8BDate=2025-122026.05 | — | — | — | — | 40.7 | |
| OpenCUA-32BDate=2025-082026.05 | — | — | — | — | 26.3 | |
| OpenCUA-7BDate=2025-082026.05 | — | — | — | — | 25.5 | |
| OS-Atlas-Base-7BDate=2024-102026.05 | — | — | — | — | 9 | |
| Phi-Ground-4B-16CDate=2025-072026.05 | — | — | — | — | 24.5 | |
| Phi-Ground-Any-4BDate=2026-052026.05 | — | — | — | — | 15.8 | |
| Phi-Ground-Any-4B + APP data finetunedDate=2026-05, Configuration=APP data finetuned2026.05 | — | — | — | — | 29.7 | |
| Uground-V1-2BDate=2024-102026.05 | — | — | — | — | 12.8 | |
| Uground-V1-7BDate=2024-102026.05 | — | — | — | — | 12.9 | |
| UI-TARS-1.5-7BDate=2025-042026.05 | — | — | — | — | 22.3 | |
| UI-Venus-Ground-72BDate=2025-082026.05 | — | — | — | — | 36.8 | |
| UI-Venus-Ground-7BDate=2025-082026.05 | — | — | — | — | 26.5 |