Web Navigation on Classifieds
13.7Success RateGPT-4o
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| GPT-4oModel=GPT-4o, Strategy=Zero-Shot2026.05 | 13.7 | 12.5 | |
| SCALEModel=Qwen2.5-VL-7B, Strategy=SCALE(ours)2026.05 | 12 | 4.1 | |
| Qwen2.5-VL-7BModel=Qwen2.5-VL-7B, Strategy=GPT Trajectory Imitation2026.05 | 10.7 | 9.2 | |
| Qwen2.5-VL-7BModel=Qwen2.5-VL-7B, Strategy=OS-Genesis2026.05 | 8.6 | 11.3 | |
| ViGoRLModel=ViGoRL, Strategy=Zero-Shot2026.05 | 6.9 | 12.4 | |
| SCALEModel=InternVL2.5-8B, Strategy=SCALE(ours)2026.05 | 6.4 | 5.5 | |
| Qwen2.5-VL-7BModel=Qwen2.5-VL-7B, Strategy=Zero-Shot2026.05 | 6 | 15.8 | |
| InternVL2.5-8BModel=InternVL2.5-8B, Strategy=Tree Search2026.05 | 4.3 | — | |
| InternVL2.5-8BModel=InternVL2.5-8B, Strategy=GPT Trajectory Imitation2026.05 | 3.4 | 16.7 | |
| InternVL2.5-8BModel=InternVL2.5-8B, Strategy=OS-Genesis2026.05 | 3 | 6.3 | |
| Qwen2.5-VL-7BModel=Qwen2.5-VL-7B, Strategy=Tree Search2026.05 | 1.7 | — | |
| SCALE-20kModel=LLaVA-NeXT-8B, Strategy=SCALE-20k2026.05 | 0.8 | 8.2 | |
| AugvisModel=Augvis, Strategy=Zero-Shot2026.05 | 0.4 | — | |
| InternVL2.5-8BModel=InternVL2.5-8B, Strategy=Zero-Shot2026.05 | 0.4 | 22.7 | |
| LLaVA-NeXT-8BModel=LLaVA-NeXT-8B, Strategy=Zero-Shot2026.05 | 0 | — |