Aesthetic Image Editing on AesRecon
65.33Win Rate vs. Poor (GPT-4o)AesFormer
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| AesFormerThinker=AesThinker2026.05 | 65.33 | 68.63 | 26.25 | 24.39 | 47.76 | 5.6 | 3.51 | |
| Nano Banana Pro (Google)Thinker=None2026.05 | 54.44 | 72.55 | 16.67 | 21.95 | 50.9 | 5.59 | 3.24 | |
| Qwen-Image-Edit-2511Thinker=None2026.05 | 16.5 | 9.8 | 7.64 | 12.2 | 46.65 | 5.44 | 3.2 | |
| Step1X-Edit-v1.1Thinker=None2026.05 | 15.28 | 11.76 | 13.84 | 13.41 | 37.14 | 5.33 | 3.37 | |
| Qwen-Image-Edit-2511Thinker=GPT-4o2026.05 | 14.62 | 23.53 | 14.4 | 14.63 | 44.16 | 5.36 | 3.22 | |
| FLUX.1 Kontext [Dev]Thinker=None2026.05 | 12.96 | 5.88 | 2.66 | 3.66 | 38.34 | 5.07 | 2.83 | |
| FLUX.1 Kontext [Dev]Thinker=GPT-4o2026.05 | 12.51 | 13.72 | 3.88 | 6.1 | 38.1 | 5.05 | 2.76 | |
| BagelThinker=None2026.05 | 12.4 | 17.65 | 7.75 | 12.2 | 37.69 | 4.94 | 2.58 | |
| FLUX.1 Kontext [Dev]Thinker=Qwen32026.05 | 11.96 | 7.84 | 2.88 | 4.88 | 38.6 | 5.08 | 2.8 | |
| Step1X-Edit-v1.1Thinker=Qwen32026.05 | 7.64 | 9.62 | 8.31 | 9.76 | 36.98 | 5.1 | 3.15 | |
| BagelThinker=GPT-4o2026.05 | 7.31 | 21.57 | 2.99 | 7.32 | 33.99 | 4.65 | 2.33 | |
| Qwen-Image-Edit-2511Thinker=Qwen32026.05 | 6.76 | 7.84 | 4.98 | 9.76 | 46.46 | 5.3 | 3.24 | |
| Step1X-Edit-v1.1Thinker=GPT-4o2026.05 | 6.2 | 15.69 | 6.76 | 8.54 | 36.62 | 5.01 | 3 | |
| BagelThinker=Qwen32026.05 | 3.43 | 19.61 | 2.32 | 6.1 | 35.18 | 4.74 | 2.39 |