Operating System GUI Agentic Reasoning on OSWorld
64.29Success RateEAPO-8B
Evaluation Results
| Method | Links | |
|---|---|---|
| EAPO-8BModel Category=Ours2026.05 | 64.29 | |
| EAPO-8B2026.05 | 64.29 | |
| Claude-4-Sonnet-0929Model Category=Closed-source Models2026.05 | 62.88 | |
| Claude-4-Sonnet-0929Category=Closed-source2026.05 | 62.88 | |
| EAPO-4BModel Category=Ours2026.05 | 57.89 | |
| EAPO-4B2026.05 | 57.89 | |
| LAMERCategory=Training Method2026.05 | 55.6 | |
| DeepMiner-Mano-72BModel Category=Closed-source Models2026.05 | 53.88 | |
| DeepMiner-Mano-72BCategory=Closed-source2026.05 | 53.88 | |
| UI-TARS-2-2509Model Category=Closed-source Models2026.05 | 53.11 | |
| UI-TARS-2-2509Category=Closed-source2026.05 | 53.11 | |
| GiGPOCategory=Training Method2026.05 | 50.88 | |
| EAPO-1.7B/2BModel Category=Ours2026.05 | 50.34 | |
| EAPO-1.7B/2B2026.05 | 50.34 | |
| DAPOCategory=Training Method2026.05 | 47.9 | |
| DART-GUI-7BModel Category=Open-source Models2026.05 | 42.13 | |
| DART-GUI-7BCategory=Open-source2026.05 | 42.13 | |
| Qwen3-VL-32BModel Category=Open-source Models2026.05 | 41 | |
| GRPOCategory=Training Method2026.05 | 40.36 | |
| Seed1.5-VL-250717Model Category=Closed-source Models2026.05 | 40.18 | |
| Seed1.5-VL-250717Category=Closed-source2026.05 | 40.18 | |
| DeepMiner-Mano-7BModel Category=Closed-source Models2026.05 | 40.16 | |
| Qwen3-VL-235B-A22BModel Category=Open-source Models2026.05 | 38.1 | |
| Qwen3-VL-235B-A22BCategory=Open-source2026.05 | 38.1 | |
| Claude-3.7-SonnetModel Category=Closed-source Models2026.05 | 35.83 | |
| OpenCUA-32BModel Category=Open-source Models2026.05 | 34.79 | |
| GUI-Owl-7BModel Category=Open-source Models2026.05 | 34.79 | |
| OpenCUA-32BCategory=Open-source2026.05 | 34.79 | |
| GUI-Owl-7BCategory=Open-source2026.05 | 34.79 | |
| OpenAI CUAModel Category=Closed-source Models2026.05 | 31.3 | |
| ARPOModel Category=Open-source Models2026.05 | 29.9 | |
| ARPOCategory=Open-source2026.05 | 29.9 | |
| TianXi-Action-7BModel Category=Closed-source Models2026.05 | 29.81 | |
| TianXi-Action-7BCategory=Closed-source2026.05 | 29.81 | |
| OpenCUA-7BModel Category=Open-source Models2026.05 | 28.2 | |
| UI-TARS-7BModel Category=Open-source Models2026.05 | 27.52 | |
| UI-TARS-7BCategory=Open-source2026.05 | 27.52 | |
| UI-TARS-72B-dpoModel Category=Open-source Models2026.05 | 26.84 | |
| OpenAI CUA o3Model Category=Closed-source Models2026.05 | 23 | |
| OpenAI CUA o3Category=Closed-source2026.05 | 23 | |
| ZeroGUIModel Category=Open-source Models2026.05 | 20.2 | |
| ZeroGUICategory=Open-source2026.05 | 20.2 |