GUI Navigation on AITW
67.3Overall Success RateQwen-GUI
Evaluation Results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Qwen-GUI2024.06 | 67.3 | 70.3 | 61.2 | 71.6 | 66.1 | — | — | — | — | — | — | — | — | |
| Qwen-VL2024.06 | 65.8 | 70.4 | 57.8 | 70.1 | 64.7 | — | — | — | — | — | — | — | — | |
| SeeClick2024.06 | 60.6 | 66.4 | 54.9 | 63.5 | 57.6 | — | — | — | — | — | — | — | — | |
| NaviMastertraining_data=GUI + Embodied2025.08 | 59.72 | — | — | — | — | 74.92 | — | — | — | — | — | — | — | |
| MiniCPM-GUI2024.06 | 58.4 | 62.3 | 46.5 | 67.3 | 57.5 | — | — | — | — | — | — | — | — | |
| NaviMastertraining_data=GUI only2025.08 | 58.05 | — | — | — | — | 71.31 | — | — | — | — | — | — | — | |
| OmniParser (w. LS + ID)Model Source=Close-Source Model, Evaluation Protocol=Zero-shot2026.06 | 57.7 | 57.8 | 51.6 | 77.4 | 52.9 | — | 48.3 | — | — | — | — | — | — | |
| GUI-R1-7B2025.08 | 55.31 | — | — | — | — | 62.94 | — | — | — | — | — | — | — | |
| UI-shift2025.08 | 54.38 | — | — | — | — | 74.44 | — | — | — | — | — | — | — | |
| Aguvis2025.08 | 53.06 | — | — | — | — | 69.57 | — | — | — | — | — | — | — | |
| GPT-4V + historyModel Source=Close-Source Model, Evaluation Protocol=Zero-shot2026.06 | 52.96 | 46.14 | 49.18 | 78.29 | 48.18 | — | 43.01 | — | — | — | — | — | — | |
| GUICrafter–3BModel Source=Open-Source Model, Evaluation Protocol=Zero-shot2026.06 | 50.89 | 54.29 | 51.86 | 61.46 | 40.25 | — | 46.57 | — | — | — | — | — | — | |
| GUICrafter–3BModel Source=Open-Source Model, Evaluation Protocol=Zero-shot, Training Stage=only Stage22026.06 | 49.92 | 52.11 | 52.36 | 58.84 | 41.28 | — | 45 | — | — | — | — | — | — | |
| NaviMastertraining_data=Embodied only2025.08 | 49.17 | — | — | — | — | 74.95 | — | — | — | — | — | — | — | |
| MiniCPM-V2024.06 | 48.9 | 50.2 | 45.1 | 56.2 | 44 | — | — | — | — | — | — | — | — | |
| GoClick-L w/ Intent Gnd.Planner=GPT-4o, Grounding Model=GoClick-L w/ Intent Gnd.2026.04 | 48.9 | 50 | 49.7 | 54.4 | 44.5 | — | 45.9 | 57.4 | 59 | 57.1 | 69.7 | 60.5 | 59.7 | |
| Qwen2.5VL-7Bvariant=*2025.08 | 48.81 | — | — | — | — | 62.51 | — | — | — | — | — | — | — | |
| UI-AGILE2025.08 | 48.66 | — | — | — | — | 75.72 | — | — | — | — | — | — | — | |
| GUICrafter–3BModel Source=Open-Source Model, Evaluation Protocol=Zero-shot, Reward Strategy=Binary Reward2026.06 | 48.51 | 52.28 | 51.15 | 57.68 | 36.99 | — | 44.44 | — | — | — | — | — | — | |
| GoClick-L w/ Func. Gnd.Planner=GPT-4o, Grounding Model=GoClick-L w/ Func. Gnd.2026.04 | 47.6 | 47.1 | 47.9 | 53.5 | 44 | — | 45.7 | 56.9 | 51.2 | 54.1 | 67.4 | 59.7 | 57.5 | |
| GoClick-L w/ Intent Gnd.Planner=Gemini-2-Flash-Exp, Grounding Model=GoClick-L w/ Intent Gnd.2026.04 | 47.2 | 44.1 | 49.5 | 59.7 | 39.8 | — | 43.1 | 48.1 | 52.9 | 54.6 | 67.4 | 54.8 | 54 | |
| infiGUI-3B2025.08 | 46.51 | — | — | — | — | 75.58 | — | — | — | — | — | — | — | |
| GoClick-L w/ Func. Gnd.Planner=Gemini-2-Flash-Exp, Grounding Model=GoClick-L w/ Func. Gnd.2026.04 | 46.4 | 40.9 | 48.5 | 59.4 | 40 | — | 43.2 | 48.4 | 47.4 | 52.9 | 66.8 | 55 | 52.5 | |
| Qwen2.5VL-7B2025.08 | 45.8 | — | — | — | — | 65.99 | — | — | — | — | — | — | — | |
| Fuyu-GUI2024.06 | 45.5 | 50.9 | 41.6 | 45.7 | 43.8 | — | — | — | — | — | — | — | — | |
| GUI-R1-3BModel Source=Open-Source Model, Evaluation Protocol=Zero-shot2026.06 | 43.6 | 44.87 | 39.97 | 54.06 | 35.6 | — | 43.48 | — | — | — | — | — | — | |
| Fuyu-8B2024.06 | 43.5 | 45.9 | 40 | 47.2 | 40.8 | — | — | — | — | — | — | — | — | |
| GPT-4o + SoMPlanner=GPT-4o + SoM, Grounding Model=None2026.04 | 42.1 | 43.2 | 41.4 | 53 | 39.2 | — | 33.7 | 37.2 | 53.8 | 51.8 | 63.2 | 51 | 50.4 | |
| OS-Atlas-7B2025.08 | 41.38 | — | — | — | — | 64.89 | — | — | — | — | — | — | — | |
| PalM2-CoTModel Source=Close-Source Model, Evaluation Protocol=Zero-shot2026.06 | 39.6 | — | — | — | — | — | — | — | — | — | — | — | — | |
| ShowUI-2BModel Source=Open-Source Model, Evaluation Protocol=Zero-shot2026.06 | 35.9 | 47.7 | 42 | 20.1 | 37.4 | — | 32.1 | — | — | — | — | — | — | |
| Gemini-2-Flash-Exp + SoMPlanner=Gemini-2-Flash-Exp + SoM, Grounding Model=None2026.04 | 34.8 | 33.9 | 33.9 | 48.5 | 27.9 | — | 29.9 | 32.2 | 41.4 | 30.1 | 56.8 | 37.8 | 38.3 | |
| GUICrafter–3BModel Source=Open-Source Model, Evaluation Protocol=Zero-shot, Training Stage=only Stage12026.06 | 31.53 | 34.07 | 32.73 | 36.26 | 24.5 | — | 30.09 | — | — | — | — | — | — | |
| Gemini-2-Flash-ExpPlanner=Gemini-2-Flash-Exp, Grounding Model=None2026.04 | 29.5 | 28.5 | 30.3 | 41.9 | 20.2 | — | 26.4 | 18.2 | 26.9 | 22.9 | 29 | 22.7 | 23.6 | |
| Qwen2.5-VL-3BModel Source=Open-Source Model, Evaluation Protocol=Zero-shot2026.06 | 28.14 | 28.58 | 22.97 | 37.28 | 24.3 | — | 27.57 | — | — | — | — | — | — | |
| GPT-4oPlanner=GPT-4o, Grounding Model=None2026.04 | 27.2 | 32.9 | 31.9 | 44.2 | 25.9 | — | 28.2 | 24.9 | 30.1 | 27.6 | 47.1 | 30 | 29.9 | |
| GPT-4o2025.08 | 26.07 | — | — | — | — | 35.75 | — | — | — | — | — | — | — | |
| ChatGPT-CoTModel Source=Close-Source Model, Evaluation Protocol=Zero-shot2026.06 | 7.72 | 4.38 | 10.47 | 9.39 | 8.42 | — | 5.93 | — | — | — | — | — | — |