Single-step Trajectory Evaluation on RISK-Bench Single-step
99.1Easy ScoreRISK-SFT-7B
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| RISK-SFT-7BModel Category=Ours, Base Model=Qwen2.5-VL-7B-Instruct, Training Type=SFT2025.09 | 99.1 | 83.2 | 52.5 | 83.5 | |
| Qwen2.5-VL-72BModel Category=General Open-source Models2025.09 | 98.8 | 81.9 | 42.9 | 80.6 | |
| RISK-R1-7BModel Category=Ours, Base Model=Qwen2.5-VL-7B-Instruct, Training Type=RFT2025.09 | 98.8 | 90.1 | 65.5 | 88.3 | |
| GPT-4oModel Category=Commercial Models2025.09 | 98.2 | 82.9 | 46.8 | 81.5 | |
| Qwen-VL-MaxModel Category=Commercial Models2025.09 | 95.8 | 78.5 | 22.4 | 72.9 | |
| GUI-R1-7BModel Category=GUI-specific Models (RL)2025.09 | 65.4 | 45 | 9.3 | 46.3 | |
| Qwen2.5-VL-7BModel Category=General Open-source Models2025.09 | 62.3 | 45.4 | 4.8 | 43.6 | |
| GUI-R1-3BModel Category=GUI-specific Models (RL)2025.09 | 55.4 | 37.2 | 3 | 37.8 | |
| GUI-G1-3BModel Category=GUI-specific Models (RL)2025.09 | 55.1 | 38.9 | 4.5 | 38.4 | |
| UI-TARS-1.5-7BModel Category=GUI-specific Models (RL)2025.09 | 44.9 | 28.4 | 1.9 | 30.1 | |
| OS-Atlas-7BModel Category=GUI-specific Models (SFT)2025.09 | 37.3 | 27.7 | 2.4 | 26.1 | |
| InternVL3-8BModel Category=General Open-source Models2025.09 | 30.1 | 14.3 | 0 | 18.8 | |
| UI-TARS-72BModel Category=GUI-specific Models (SFT)2025.09 | 20.7 | 9.3 | 0.9 | 13 | |
| InfiGUI-R1-3BModel Category=GUI-specific Models (RL)2025.09 | 18.6 | 11.7 | 1.4 | 12.6 | |
| UI-TARS-7BModel Category=GUI-specific Models (SFT)2025.09 | 11.3 | 5.5 | 0 | 7.1 | |
| Aguvis-7BModel Category=GUI-specific Models (SFT)2025.09 | 8.3 | 4.5 | 0 | 5.3 | |
| ShowUI-2BModel Category=GUI-specific Models (SFT)2025.09 | 6.9 | 2.7 | 0 | 4.2 | |
| UI-Venus-Navi-7BModel Category=GUI-specific Models (RL)2025.09 | 0.7 | 0 | 0 | 0.3 | |
| UI-TARS-2BModel Category=GUI-specific Models (SFT)2025.09 | 0.2 | 0 | 0 | 0.1 |