GUI Navigation on GUI-Odyssey
68.32AMSWorking + Episodic Memory
Evaluation Results
| Method | Links | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Working + Episodic MemoryBackbone=UI-Venus-1.5-8B2026.05 | 68.32 | 3.57 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Working MemoryBackbone=UI-Venus-1.5-8B2026.05 | 67.69 | 2.69 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Pred. Hist. AllBackbone=UI-Venus-1.5-8B2026.05 | 66.31 | 2.33 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Text Summary MemoryBackbone=UI-Venus-1.5-8B2026.05 | 62.18 | 2.12 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Direct PromptingBackbone=Gemini-3.1-Pro2026.05 | 60.62 | 1.81 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Working + Episodic MemoryBackbone=GUI-Owl-1.5-32B2026.05 | 55.17 | 2.59 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| No HistoryBackbone=UI-Venus-1.5-8B2026.05 | 54.58 | 1.29 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Direct PromptingBackbone=GPT-5.52026.05 | 54.46 | 2.02 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Working MemoryBackbone=GUI-Owl-1.5-32B2026.05 | 51.36 | 2.33 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Working + Episodic MemoryBackbone=GUI-Owl-1.5-8B2026.05 | 49.45 | 1.71 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Working + Episodic MemoryBackbone=MAI-UI-8B2026.05 | 49.31 | 2.12 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Working MemoryBackbone=MAI-UI-8B2026.05 | 49.08 | 1.97 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Pred. Hist. AllBackbone=GUI-Owl-1.5-32B2026.05 | 49.02 | 1.55 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Working MemoryBackbone=GUI-Owl-1.5-8B2026.05 | 48.25 | 1.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Text Summary MemoryBackbone=GUI-Owl-1.5-32B2026.05 | 47.21 | 0.72 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| No HistoryBackbone=GUI-Owl-1.5-32B2026.05 | 45.73 | 0.57 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Text Summary MemoryBackbone=GUI-Owl-1.5-8B2026.05 | 45.45 | 0.62 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Pred. Hist. AllBackbone=MAI-UI-8B2026.05 | 44.79 | 1.35 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| No HistoryBackbone=GUI-Owl-1.5-8B2026.05 | 40.15 | 0.16 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Pred. Hist. AllBackbone=GUI-Owl-1.5-8B2026.05 | 38.88 | 0.47 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Text Summary MemoryBackbone=MAI-UI-8B2026.05 | 38.33 | 0.62 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| No HistoryBackbone=MAI-UI-8B2026.05 | 35.7 | 0.36 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Adaptive Multimodal Agents-Based FrameworkModel Size=30B2026.05 | — | — | 68.2 | 3.9 | 52.8 | 0 | 49.5 | 1.16 | 50.8 | 48 | 58.9 | 1.32 | 55.2 | 1.54 | 56.4 | 1.55 | |
| CogAgent2026.05 | — | — | 11.8 | — | 15.7 | — | 10.7 | — | 9.2 | — | 11.7 | — | 13.1 | — | 10.7 | — | |
| GeminiProVision2026.05 | — | — | 3.3 | — | 4 | — | 2.3 | — | 4.3 | — | 1.5 | — | 3.2 | — | 4.9 | — | |
| GPT-4V2026.05 | — | — | 18.8 | — | 23.5 | — | 20.2 | — | 19.2 | — | 16.9 | — | 13.8 | — | 19 | — | |
| PG-AgentModel Size=72B2026.05 | — | — | 48.6 | — | 61.5 | — | 47.2 | — | 35.5 | — | 46.9 | — | 52.6 | — | 47.7 | — | |
| PG-AgentModel Size=30B2026.05 | — | — | 53.5 | 0.84 | 38.8 | 0 | 35.6 | 0 | 39 | 89 | 47.9 | 0.83 | 43.7 | 0 | 43.6 | 0.42 | |
| Qwen2.5-VL-72BModel Size=72B2026.05 | — | — | 46.6 | — | 60 | — | 44 | — | 32.4 | — | 46.1 | — | 54.6 | — | 42.4 | — |