Instructive Image Editing on MagicBrush (test)
0.9281CLIP ImageUNO + OPRO
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| UNO + OPROTrainable Parameters=+0.93M, rho=322026.03 | 0.9281 | — | — | 0.0387 | 0.898 | |
| ReasonBrain2025.07 | 0.928 | 0.139 | 0.281 | 0.049 | 0.893 | |
| UNOTrainable Parameters=478.2M2026.03 | 0.9236 | — | — | 0.0575 | 0.8961 | |
| EditMGTModel Category=Universal model2025.12 | 0.911 | — | 0.301 | 0.058 | 0.881 | |
| OmniGen2Category=Unified Models, Venue=CVPR’25, Backbone=FLUX.1 Dev2026.03 | 0.905 | — | 0.306 | 0.055 | 0.889 | |
| SmartFreeEdit2025.07 | 0.905 | 0.133 | 0.27 | 0.054 | 0.887 | |
| UniRealModel Type=Universal2024.12 | 0.903 | 0.151 | 0.265 | 0.063 | 0.876 | |
| UniRealModel Category=Universal model2025.12 | 0.903 | — | 0.308 | 0.081 | 0.837 | |
| MIGE2025.07 | 0.902 | 0.131 | 0.268 | 0.056 | 0.882 | |
| VAREdit-8BModel Category=Universal model2025.12 | 0.901 | — | 0.287 | 0.083 | 0.844 | |
| ICEdit + OPROTrainable Parameters=+0.93M, rho=322026.03 | 0.9002 | — | — | 0.0781 | 0.8531 | |
| AnyEditModel Category=Task-specific2025.12 | 0.898 | — | 0.275 | 0.051 | 0.881 | |
| EMU EditModel Type=Task-specific2024.12 | 0.897 | 0.135 | 0.284 | — | — | |
| EMU EditModel Category=Task-specific2025.12 | 0.897 | — | 0.261 | 0.052 | 0.879 | |
| EMU-EditCategory=Task-specific Models, Venue=CVPR’24, Backbone=–2026.03 | 0.897 | — | 0.261 | 0.052 | 0.879 | |
| Emu Edit2025.07 | 0.897 | 0.135 | 0.261 | 0.052 | 0.879 | |
| OmniGenModel Type=Universal2024.12 | 0.895 | 0.086 | 0.274 | 0.076 | 0.862 | |
| SmartEdit2025.07 | 0.895 | 0.119 | 0.262 | 0.094 | 0.82 | |
| CARE-EditCategory=Unified Models, Venue=–, Backbone=FLUX.1 Dev2026.03 | 0.894 | — | 0.324 | 0.052 | 0.885 | |
| AnyEditCategory=Unified Models, Venue=CVPR’25, Backbone=SD1.52026.03 | 0.892 | — | 0.273 | 0.057 | 0.877 | |
| GPT Image 1.52025.07 | 0.892 | 0.132 | 0.266 | 0.055 | 0.885 | |
| Qwen-Image-Edit2025.07 | 0.89 | 0.13 | 0.264 | 0.057 | 0.881 | |
| Flux Kontext2025.07 | 0.887 | 0.126 | 0.263 | 0.06 | 0.875 | |
| Bagel-Thinking2025.07 | 0.885 | 0.124 | 0.262 | 0.062 | 0.872 | |
| PixWizardModel Type=Universal2024.12 | 0.884 | 0.124 | 0.308 | 0.081 | 0.837 | |
| PixWizardModel Category=Universal model2025.12 | 0.884 | — | 0.265 | 0.063 | 0.876 | |
| PixWizard2025.07 | 0.884 | 0.124 | 0.265 | 0.063 | 0.876 | |
| MagicBrushModel Type=Task-specific2024.12 | 0.883 | 0.123 | 0.261 | 0.058 | 0.871 | |
| MagicBrushModel Category=Task-specific2025.12 | 0.883 | — | 0.261 | 0.058 | 0.871 | |
| MagicBrush2025.07 | 0.883 | 0.123 | 0.261 | 0.058 | 0.871 | |
| Bagel2025.07 | 0.882 | 0.121 | 0.26 | 0.065 | 0.868 | |
| ChatGPT-4o2025.07 | 0.88 | 0.122 | 0.259 | 0.064 | 0.865 | |
| ACE++ + OPROTrainable Parameters=+0.93M, rho=322026.03 | 0.8749 | — | — | 0.1114 | 0.7767 | |
| InsertAnything + OPROTrainable Parameters=+0.93M, rho=322026.03 | 0.8735 | — | — | 0.1269 | 0.8009 | |
| InsertAnythingTrainable Parameters=37.5M2026.03 | 0.8722 | — | — | 0.1327 | 0.7917 | |
| ICEditTrainable Parameters=22.4M2026.03 | 0.8703 | — | — | 0.1189 | 0.7706 | |
| UltraEditModel Category=Task-specific2025.12 | 0.868 | — | — | 0.088 | 0.792 | |
| UltraEdit2025.07 | 0.868 | — | — | 0.088 | 0.792 | |
| EditARModel Category=Universal model2025.12 | 0.867 | — | — | 0.103 | 0.804 | |
| ACE++Trainable Parameters=76.6M2026.03 | 0.8658 | — | — | 0.1215 | 0.7394 | |
| InstructPix2PixModel Type=Task-specific2024.12 | 0.837 | 0.115 | 0.245 | 0.093 | 0.767 | |
| InstructPix2PixModel Category=Task-specific2025.12 | 0.837 | — | 0.245 | 0.093 | 0.767 | |
| InstructPix2PixCategory=Task-specific Models, Venue=CVPR’23, Backbone=SD1.52026.03 | 0.837 | — | 0.245 | 0.093 | 0.767 | |
| InstructPix2Pix2025.07 | 0.837 | 0.115 | 0.245 | 0.093 | 0.767 | |
| ACE (ACE++)Category=Unified Models, Venue=ICLR’26, Backbone=FLUX.1 Fill2026.03 | 0.818 | — | 0.268 | 0.042 | 0.823 | |
| UltraEditModel Type=Task-specific2024.12 | 0.752 | 0.121 | 0.263 | 0.077 | 0.664 | |
| Null-Text Inv.Model Category=Task-specific2025.12 | 0.752 | — | 0.263 | 0.077 | 0.664 | |
| Null-TextCategory=Task-specific Models, Venue=CVPR’23, Backbone=SD1.52026.03 | 0.752 | — | 0.263 | 0.077 | 0.664 | |
| MGIE2025.07 | 0.745 | 0.116 | 0.251 | 0.162 | 0.577 | |
| FLUX.1 FillCategory=Unified Models, Venue=HuggingFace, Backbone=FLUX.1 Fill2026.03 | 0.725 | — | 0.235 | 0.208 | 0.661 | |
| PnPModel Type=Task-specific2024.12 | 0.568 | 0.025 | — | — | — | |
| PnPModel Category=Task-specific2025.12 | 0.568 | — | 0.101 | 0.289 | 0.22 | |
| PnPCategory=Task-specific Models, Venue=CVPR’23, Backbone=SD1.52026.03 | 0.568 | — | 0.101 | 0.289 | 0.22 |