GUI Agent on AITZ
69.4SRM2-Miner-7B
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| M2-Miner-7BTraining Data=Open-source Models W/ Auto-Mined Datasets2026.02 | 69.4 | 81.3 | — | — | |
| M2-Miner-3BTraining Data=Open-source Models W/ Auto-Mined Datasets2026.02 | 66.6 | 78.6 | — | — | |
| UI-TARS-7B*Training Data=Open-source Models W/ Private Human-Annotated Datasets2026.02 | 65.8 | 80.4 | — | — | |
| GUI-Owl-7BTraining Data=Open-source Models W/ Auto-Mined Datasets2026.02 | 65.1 | 78.9 | — | — | |
| Learn from AITZBase Model=OS-Atlas-Pro-7B, Learning Strategy=Supervised Fine-Tuning, Training Source=AITZ2026.01 | 59.92 | — | 75.63 | 613 | |
| OS-Atlas-7BTraining Data=Open-source Models W/ Public Datasets2026.02 | 58.5 | 74.1 | — | — | |
| Agent-DiceBase Model=OS-Atlas-Pro-7B, Learning Strategy=Geometric Consensus & Curvature Importance Weighting2026.01 | 57.1 | — | 74.72 | 237 | |
| Qwen2.5-VL-7BTraining Data=Open-source Models W/ Public Datasets2026.02 | 54.6 | 78.4 | — | — | |
| InfiGUI-R1Training Data=Open-source Models W/ Public Datasets2026.02 | 52.9 | 70.7 | — | — | |
| Qwen2.5-VL-3BTraining Data=Open-source Models W/ Public Datasets2026.02 | 52.7 | 75.1 | — | — | |
| GUI-R1-7BTraining Data=Open-source Models W/ Public Datasets2026.02 | 50.5 | 56.8 | — | — | |
| Zero-ShotBase Model=OS-Atlas-Pro-7B, Learning Strategy=Zero-Shot2026.01 | 45.08 | — | 63.41 | 20 | |
| CL from all threeBase Model=OS-Atlas-Pro-7B, Learning Strategy=Continual Learning, Training Source=AITZ, AndroidControl, GUI-Odyssey2026.01 | 42.05 | — | 65.78 | 4 | |
| Learn from GUI-OdysseyBase Model=OS-Atlas-Pro-7B, Learning Strategy=Supervised Fine-Tuning, Training Source=GUI-Odyssey2026.01 | 37.67 | — | 63.16 | 20 | |
| CL from AITZ and AndroidControlBase Model=OS-Atlas-Pro-7B, Learning Strategy=Continual Learning, Training Source=AITZ, AndroidControl2026.01 | 37.41 | — | 65.81 | 0 | |
| GPT-4oTraining Data=Closed-source Models2026.02 | 35.3 | 70 | — | — | |
| OdysseyAgentTraining Data=Open-source Models W/ Public Datasets2026.02 | 31.6 | 59.2 | — | — | |
| Learn from AndroidControlBase Model=OS-Atlas-Pro-7B, Learning Strategy=Supervised Fine-Tuning, Training Source=AndroidControl2026.01 | 30.85 | — | 61.34 | 0 | |
| Aguvis-7BTraining Data=Open-source Models W/ Public Datasets2026.02 | 19 | 35.7 | — | — | |
| OS-Genesis-7BTraining Data=Open-source Models W/ Auto-Mined Datasets2026.02 | 8.5 | 20 | — | — |