Visual Navigation on Visual Navigation (level-3)
93.8Pass@1VisuoThink
Evaluation Results
| Method | Links | |
|---|---|---|
| VisuoThinkModel=GPT-4o2025.04 | 93.8 | |
| VisuoThinkModel=Claude-3.5-sonnet2025.04 | 93.8 | |
| VisuoThinkModel=Qwen2-VL-72B-Instruct2025.04 | 81.3 | |
| VisuoThink w/o rollout searchModel=GPT-4o2025.04 | 81.2 | |
| VisuoThink w/o rollout searchModel=Claude-3.5-sonnet2025.04 | 81.2 | |
| VoT + ExecuterModel=Claude-3.5-sonnet2025.04 | 68.8 | |
| VoT + ExecuterModel=GPT-4o2025.04 | 62.5 | |
| VoTModel=Claude-3.5-sonnet2025.04 | 56.3 | |
| VisuoThink w/o rollout searchModel=Qwen2-VL-72B-Instruct2025.04 | 50 | |
| CoTModel=Claude-3.5-sonnet2025.04 | 37.5 | |
| VoTModel=GPT-4o2025.04 | 25 | |
| VoT + ExecuterModel=Qwen2-VL-72B-Instruct2025.04 | 25 | |
| CoTModel=GPT-4o2025.04 | 18.8 | |
| CoTModel=Qwen2-VL-72B-Instruct2025.04 | 6.7 | |
| VoTModel=Qwen2-VL-72B-Instruct2025.04 | 0 |