GUI Interaction Control on GUI-Odyssey
87SRUI-TARS-7B
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| UI-TARS-7BModel Type=Open-source, Model Scale=7B2026.04 | 87 | — | 90.1 | 94.6 | — | |
| AgentCPM-GUI-8BModel Type=Open-source, Model Scale=8B2026.04 | 75 | — | — | 90.8 | — | |
| UI-S1-7BModel Type=Open-source, Model Scale=7B2026.04 | 59.5 | — | 61.7 | 76.3 | — | |
| UI-Copilot-7BModel Type=Open-source, Model Scale=7B2026.04 | 57.2 | — | 63.8 | 74.5 | — | |
| WebChain-LCRL-7B +SGRL+CoT-SFTTraining Paradigm=Trained on WebChain2026.03 | 54.8 | 88.7 | 57.9 | — | — | |
| SeeClickModel Type=Open-source2026.04 | 53.9 | — | 52.4 | 71 | — | |
| WebChain-LCRL-7B +CoT-SFTTraining Paradigm=Trained on WebChain2026.03 | 52.7 | 86.1 | 56.3 | — | — | |
| Qwen2.5VL-7BModel Type=Open-source, Model Scale=7B2026.04 | 52.4 | — | 56.3 | 67.4 | — | |
| WebChain-LCRL-7BTraining Paradigm=Trained on WebChain2026.03 | 50.1 | 84.4 | 54 | — | — | |
| WebChain-LCRL-3B +SGRL+CoT-SFTTraining Paradigm=Trained on WebChain2026.03 | 49.2 | 85.5 | 54.4 | — | — | |
| WebChain-LCRL-3B +CoT-SFTTraining Paradigm=Trained on WebChain2026.03 | 46.6 | 84.2 | 50.1 | — | — | |
| WebChain-LCRL-3BTraining Paradigm=Trained on WebChain2026.03 | 45.7 | 81.4 | 48.7 | — | — | |
| UI-R1-7BModel Type=Open-source, Model Scale=7B2026.04 | 43.5 | — | 41.3 | 67.1 | — | |
| PAL-UI-7BModel Type=Open-source, Model Scale=7B2026.04 | 41.7 | — | 46.8 | 65.1 | — | |
| GUI-R1-3BTraining Paradigm=Trained on Other Datasets2026.03 | 41.3 | 54.8 | 41.5 | — | — | |
| GUI-R1-3BModel Type=Open-source, Model Scale=3B2026.04 | 41.3 | — | 41.5 | 54.8 | — | |
| GUI-R1-7BTraining Paradigm=Trained on Other Datasets2026.03 | 38.8 | 65.5 | 43.6 | — | — | |
| GUI-R1-7BModel Type=Open-source, Model Scale=7B2026.04 | 38.8 | — | 43.6 | 65.5 | — | |
| UI-AGILE-7BModel Type=Open-source, Model Scale=7B2026.04 | 37 | — | — | — | — | |
| NaviMaster-7BModel Type=Open-source, Model Scale=7B2026.04 | 36.9 | — | — | 64.4 | — | |
| Qwen2.5-VL-7BTraining Paradigm=Zero Shot2026.03 | 34.7 | 55.1 | 38.1 | — | — | |
| PAL-UI-3BModel Type=Open-source, Model Scale=3B2026.04 | 34.6 | — | 36.9 | 56.7 | — | |
| UI-R1-3BModel Type=Open-source, Model Scale=3B2026.04 | 32.5 | — | 34.5 | 52.2 | — | |
| OS-Atlas-7BModel Type=Open-source, Model Scale=7B2026.04 | 27 | — | 39.7 | 60.4 | — | |
| Qwen2.5-VL-3BTraining Paradigm=Zero Shot2026.03 | 26.9 | 37.9 | 27.1 | — | — | |
| Qwen2.5VL-3BModel Type=Open-source, Model Scale=3B2026.04 | 26.7 | — | 26.5 | 37.4 | — | |
| OS-Atlas-4BModel Type=Open-source, Model Scale=4B2026.04 | 20.3 | — | 34.6 | 49.6 | — | |
| Aguvis-7BModel Type=Open-source, Model Scale=7B2026.04 | 13.5 | — | — | 26.7 | — | |
| OS-Genesis-7BModel Type=Open-source, Model Scale=7B2026.04 | 3.6 | — | — | 11.7 | — | |
| GPT-4oModel Type=Closed-source2026.04 | 3.3 | — | 0 | 34.3 | — | |
| Claude-CUModel Type=Closed-source2026.04 | 3.1 | — | 0 | 60.9 | — | |
| AgentCPM-GUI-8BOnline Data=N, Training Steps=>470K, Training Trajectories=>55K2026.04 | — | — | — | 90.82 | 74.84 | |
| GLM-4.1V-ThinkingOnline Data=–, Training Steps=–, Training Trajectories=–2026.04 | — | — | — | 72.76 | 42.57 | |
| GLM-4.5V (106B)Online Data=–, Training Steps=–, Training Trajectories=–2026.04 | — | — | — | 75.33 | 48.9 | |
| GUI-Owl-7BOnline Data=Y, Training Steps=–, Training Trajectories=–2026.04 | — | — | — | 81.58 | 65.22 | |
| Qwen2.5-VL-7BOnline Data=–, Training Steps=–, Training Trajectories=–2026.04 | — | — | — | 61.89 | 47.92 | |
| SOLAR-RLOnline Data=N, Training Steps=94K, Training Trajectories=15K2026.04 | — | — | — | 87.6 | 68.2 | |
| UI-TARS-72B-SFTOnline Data=Y, Training Steps=–, Training Trajectories=145K2026.04 | — | — | — | 89.8 | 72.27 | |
| UI-TARS-7B-SFTOnline Data=Y, Training Steps=–, Training Trajectories=145K2026.04 | — | — | — | 86.94 | 68.82 | |
| UI-Venus-Navi-7BOnline Data=N, Training Steps=350K, Training Trajectories=–2026.04 | — | — | — | 87.3 | 71.09 |