Image Editing on GRADE (test)
87.5Physics ScoreGPT-Image-2
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| GPT-Image-2Access=Closed-Source2026.07 | 87.5 | 79.2 | 92.9 | 79.8 | 84.8 | 94.5 | 87.4 | 85.2 | 92.5 | 91.1 | 87.5 | |
| Nano Banana ProAccess=Closed-Source2026.07 | 84 | 75.7 | 85.6 | 75.9 | 82.8 | 89.5 | 81.7 | 80.5 | 89 | 90.1 | 82.9 | |
| Nano Banana 2Access=Closed-Source2026.07 | 77.2 | 68.8 | 83.2 | 74.7 | 76.7 | 92.6 | 75.5 | 76.2 | 83.2 | 78.1 | 79.1 | |
| Seedream 5.0Access=Closed-Source2026.07 | 74.1 | 72.9 | 79.7 | 67.4 | 62.6 | 79.5 | 66.2 | 71.5 | 82.6 | 74.8 | 73.7 | |
| OursAccess=Open-Source2026.07 | 61.6 | 68.9 | 58.2 | 52.1 | 46 | 59.1 | 57.5 | 66.6 | 65 | 59.1 | 58.7 | |
| GPT-Image-1.5Access=Closed-Source2026.07 | 60.8 | 67.8 | 72.9 | 58.5 | 62.3 | 72.1 | 53.8 | 77.1 | 74.2 | 69 | 66.5 | |
| GPT-Image-1.0Access=Closed-Source2026.07 | 55.4 | 56.7 | 61.9 | 43.3 | 51.5 | 57.8 | 44.5 | 58.4 | 60.2 | 58 | 54.2 | |
| FLUX.2 MaxAccess=Closed-Source2026.07 | 54.7 | 56 | 63.2 | 45.4 | 38.6 | 53 | 64.2 | 60.6 | 69.7 | 57.8 | 55.7 | |
| Seedream 4.5Access=Closed-Source2026.07 | 53.6 | 54.4 | 57.4 | 40.5 | 40.7 | 55.8 | 35.6 | 43.2 | 60.2 | 44.9 | 49.7 | |
| DreamOmniAccess=Open-Source2026.07 | 40.6 | 52.7 | 50.6 | 40.1 | 47.4 | 44.9 | 35.5 | 35.9 | 47 | 48.9 | 44.3 | |
| FLUX.2 devAccess=Open-Source2026.07 | 34.3 | 54.3 | 34 | 37 | 34.8 | 32.8 | 40.4 | 42.9 | 37.1 | 36.7 | 37.6 | |
| Step-1X (think)Access=Open-Source2026.07 | 30.9 | 45 | 34 | 29.3 | 31.7 | 27.5 | 49.8 | 48.7 | 38.1 | 23.7 | 34.3 | |
| BAGELAccess=Open-Source2026.07 | 27.7 | 44.3 | 35.4 | 34.3 | 34.4 | 27.2 | 38.5 | 35.2 | 34 | 30.7 | 33.7 | |
| Qwen-Edit-2511Access=Open-Source2026.07 | 27.7 | 45.1 | 13.7 | 39.5 | 19.8 | 28.6 | 36.9 | 33 | 27.5 | 23.9 | 30 | |
| OmniGenAccess=Open-Source2026.07 | 19.5 | 44 | 15.7 | 15.3 | 30 | 11.5 | 24.8 | 23.4 | 21.2 | 31.8 | 21.1 |