Web Navigation on Multimodal-Mind2Web Cross-Website
60.5Element AccuracyExplorer-7B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Explorer-7BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | 60.5 | 90.7 | 56.7 | |
| AgentTrek-7BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | 57.6 | 88.1 | 51.4 | |
| Explorer-7BLearning Paradigm=Supervised Fine-Tuning2026.05 | 56.3 | 89.7 | 52 | |
| Explorer-4BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | 55.6 | 89.5 | 51.4 | |
| Skill-CMIBLearning Paradigm=Skill Based2026.05 | 55.4 | 60.8 | 34.1 | |
| Explorer-4BLearning Paradigm=Supervised Fine-Tuning2026.05 | 49.1 | 87.2 | 45 | |
| Qwen2.5-VL-7B-Inst.Learning Paradigm=Skill Based2026.05 | 44.7 | 57.5 | 25.5 | |
| SeeActLearning Paradigm=In-Context Learning2026.05 | 38 | 67.8 | 32.4 | |
| Scribe-Agent-L.-32BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | 34.1 | 52.7 | 32.5 | |
| GPT-4Learning Paradigm=In-Context Learning2026.05 | 30.2 | 61 | 27 | |
| SeeClick-9.6BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | 21.9 | 82.9 | 18.8 | |
| ClassificationLearning Paradigm=Supervised Fine-Tuning2026.05 | 21.6 | — | — | |
| MiniCPM-GUI-3.1BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | 20.3 | 81.7 | 17.3 | |
| GPT-3.5Learning Paradigm=In-Context Learning2026.05 | 14.9 | 56.5 | 14.1 | |
| GenerationLearning Paradigm=Supervised Fine-Tuning2026.05 | 13.9 | 44.7 | 11 | |
| EDGE-9.6BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | — | — | 21.1 |