Web Navigation on Multimodal-Mind2Web Cross-Task
60.8Element AccuracyAgentTrek-7B
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| AgentTrek-7BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | 60.8 | 88.9 | 55.7 | |
| Skill-CMIBLearning Paradigm=Skill Based2026.05 | 57 | 59.8 | 39 | |
| Explorer-7BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | 56.5 | 90.3 | 53.2 | |
| Explorer-4BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | 53.4 | 88.1 | 50.7 | |
| Qwen2.5-VL-7B-Inst.Learning Paradigm=Skill Based2026.05 | 51.9 | 63.5 | 33.4 | |
| Explorer-7BLearning Paradigm=Supervised Fine-Tuning2026.05 | 51.8 | 88 | 48.3 | |
| Explorer-4BLearning Paradigm=Supervised Fine-Tuning2026.05 | 48.1 | 88 | 44.8 | |
| SeeActLearning Paradigm=In-Context Learning2026.05 | 46.4 | 73.4 | 40.2 | |
| GPT-4Learning Paradigm=In-Context Learning2026.05 | 40.8 | 63.1 | 32.3 | |
| Scribe-Agent-L.-32BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | 38 | 52.9 | 35.6 | |
| ClassificationLearning Paradigm=Supervised Fine-Tuning2026.05 | 26.8 | — | — | |
| SeeClick-9.6BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | 26.3 | 86.2 | 23.7 | |
| MiniCPM-GUI-3.1BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | 23.8 | 86.8 | 20.8 | |
| GenerationLearning Paradigm=Supervised Fine-Tuning2026.05 | 20.2 | 52 | 17.5 | |
| GPT-3.5Learning Paradigm=In-Context Learning2026.05 | 19.4 | 59.2 | 16.8 | |
| EDGE-9.6BLearning Paradigm=Data Synthesis+Supervised Fine-Tuning2026.05 | — | — | 30 |