GUI Navigation on WebVoyager
88.89Success Rate (Allrecipes)Ovis2.5S-RLOO
Evaluation Results
| Method | Links | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Ovis2.5S-RLOOTraining=Step-wise RLOO2026.02 | 88.89 | 85.37 | 67.44 | 72.09 | 76.19 | 93.02 | 83.33 | 70.45 | 82.93 | 70.73 | 62.79 | 80.43 | 77.82 | — | |
| Claude 4 sonnet CUMode=Computer Use2026.02 | 84.44 | 75.61 | 69.77 | 81.4 | 85.71 | 95.35 | 85.71 | 86.36 | 87.8 | 78.05 | 69.77 | 84.78 | 82.1 | — | |
| Ovis2.5S-GRPOTraining=Step-wise GRPO2026.02 | 82.22 | 78.05 | 62.79 | 76.74 | 80.95 | 90.7 | 95.24 | 65.91 | 90.24 | 87.8 | 60.47 | 76.09 | 78.79 | — | |
| Ovis2.5S-RF++Training=Step-wise REINFORCE++2026.02 | 77.78 | 70.73 | 69.77 | 83.72 | 80.95 | 81.4 | 92.86 | 72.73 | 75.61 | 80.49 | 72.09 | 80.43 | 78.21 | — | |
| Claude 3.7 sonnet CUMode=Computer Use2026.02 | 75.56 | 75.61 | 62.79 | 86.05 | 69.05 | 88.37 | 78.57 | 75 | 68.29 | 85.37 | 72.09 | 82.61 | 76.65 | — | |
| Ovis2.5SFTTraining=Supervised Fine-Tuning2026.02 | 66.67 | 70.73 | 58.14 | 79.07 | 78.57 | 88.37 | 69.05 | 56.82 | 70.73 | 60.98 | 44.19 | 80.43 | 68.68 | — | |
| GPT-4o2026.02 | 55.56 | 53.66 | 55.81 | 60.47 | 54.76 | 81.4 | 64.29 | 43.18 | 58.54 | 56.1 | 41.86 | 65.22 | 57.59 | — | |
| Qwen3-VL-32B2026.02 | 37.78 | 26.83 | 46.51 | 48.84 | 54.76 | 79.07 | 64.29 | 54.55 | 65.85 | 56.1 | 18.6 | 34.78 | 48.83 | — | |
| Qwen3-VL-8B2026.02 | 35.56 | 46.34 | 34.88 | 34.88 | 40.48 | 67.44 | 61.9 | 27.27 | 48.78 | 60.98 | 32.56 | 50 | 44.94 | — | |
| UI-TARS 1.5-7B2026.02 | 26.67 | 29.27 | 48.84 | 53.49 | 61.9 | 60.47 | 66.67 | 38.64 | 41.46 | 58.54 | 46.51 | 50 | 48.44 | — | |
| UI-TARS-7B2026.02 | 17.78 | 31.71 | 16.28 | 20.93 | 40.48 | 58.14 | 71.43 | 45.45 | 31.71 | 17.07 | 41.86 | 39.13 | 35.99 | — | |
| UI-TARS-72B2026.02 | 15.56 | 39.02 | 46.51 | 65.12 | 40.48 | 72.09 | 59.52 | 47.73 | 43.9 | 26.83 | 25.58 | 47.83 | 44.16 | — | |
| ALMABackbone=Qwen3.5-plus, Execution Model Category=Automatic Design2026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | 60.06 | |
| AUTOMMEMOBackbone=Qwen3.5-plus, Execution Model Category=Automatic Design2026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | 65.04 | |
| G-MemoryBackbone=Qwen3.5-plus, Execution Model Category=Text-based2026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | 59.68 | |
| IntentCUA2026.02 | — | — | — | — | — | — | — | — | — | — | — | — | — | 71.6 | |
| M2Backbone=Qwen3.5-plus, Execution Model Category=Multimodal-based2026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | 60.73 | |
| NoMemoryBackbone=Qwen3.5-plus, Execution Model Category=None2026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | 58.23 | |
| ReasoningBankBackbone=Qwen3.5-plus, Execution Model Category=Text-based2026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | 59.9 | |
| TrajectoryRetrievalBackbone=Qwen3.5-plus, Execution Model Category=Text-based2026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | 59.08 | |
| UFO22026.02 | — | — | — | — | — | — | — | — | — | — | — | — | — | 69 | |
| UI-TARS-1.52026.02 | — | — | — | — | — | — | — | — | — | — | — | — | — | 35.9 | |
| XSkillBackbone=Qwen3.5-plus, Execution Model Category=Multimodal-based2026.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | 52.61 |