Web Navigation on MM-Mind2Web
22.97Step Success Rate (SR)Direct Prompting
Evaluation Results
| Method | Links | |
|---|---|---|
| Direct PromptingBackbone=Gemini-3.1-Pro2026.05 | 22.97 | |
| Working + Episodic MemoryBackbone=GUI-Owl-1.5-32B2026.05 | 19.12 | |
| No HistoryBackbone=GUI-Owl-1.5-32B2026.05 | 18.98 | |
| Direct PromptingBackbone=GPT-5.52026.05 | 18.81 | |
| No HistoryBackbone=GUI-Owl-1.5-8B2026.05 | 17.78 | |
| Working MemoryBackbone=GUI-Owl-1.5-32B2026.05 | 17.71 | |
| Text Summary MemoryBackbone=GUI-Owl-1.5-32B2026.05 | 17 | |
| Pred. Hist. AllBackbone=GUI-Owl-1.5-32B2026.05 | 16.48 | |
| Working + Episodic MemoryBackbone=GUI-Owl-1.5-8B2026.05 | 16.42 | |
| Working MemoryBackbone=GUI-Owl-1.5-8B2026.05 | 15.53 | |
| Working + Episodic MemoryBackbone=MAI-UI-8B2026.05 | 14.67 | |
| Working MemoryBackbone=MAI-UI-8B2026.05 | 14.61 | |
| Text Summary MemoryBackbone=GUI-Owl-1.5-8B2026.05 | 14.14 | |
| Pred. Hist. AllBackbone=GUI-Owl-1.5-8B2026.05 | 13.56 | |
| Pred. Hist. AllBackbone=MAI-UI-8B2026.05 | 13.44 | |
| Text Summary MemoryBackbone=MAI-UI-8B2026.05 | 13.28 | |
| Working + Episodic MemoryBackbone=UI-Venus-1.5-8B2026.05 | 12.6 | |
| No HistoryBackbone=MAI-UI-8B2026.05 | 12.53 | |
| Working MemoryBackbone=UI-Venus-1.5-8B2026.05 | 11.8 | |
| Text Summary MemoryBackbone=UI-Venus-1.5-8B2026.05 | 11.66 | |
| Pred. Hist. AllBackbone=UI-Venus-1.5-8B2026.05 | 7.57 | |
| No HistoryBackbone=UI-Venus-1.5-8B2026.05 | 5.29 |