Online Web Navigation on WebVoyager (online evaluation)
51.2AmazonRInfiGUI-R1
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| RInfiGUI-R1Stage=After Preference Optimization2025.06 | 51.2 | 51.1 | 55.8 | 59.5 | 69 | 53.6 | 43.6 | 65.9 | 41.8 | 54.1 | |
| LPOStage=After Preference Optimization2025.06 | 51.2 | 60.5 | 64.3 | 54.7 | 71.4 | 56.1 | 47.5 | 57.5 | 38.6 | 57.6 | |
| Base ModelStage=After Supervised Fine-Tuning2025.06 | 40 | 58.1 | 53.4 | 38 | 54.7 | 65.8 | 33.3 | 56.5 | 41.8 | 48 | |
| RGUI-R1Stage=After Preference Optimization2025.06 | 35 | 37.2 | 27.9 | 33.3 | 57.1 | 50 | 35 | 56.5 | 15.9 | 37.5 | |
| RUI-R1Stage=After Preference Optimization2025.06 | 12.2 | 41.8 | 51.1 | 30.9 | 45.2 | 58.3 | 51.1 | 63 | 27.9 | 47.3 |