Medical Image Editing on Med-Banana-80K v1.0 (test)
8.23Mean Instruction ComplianceGPT-Image-2
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| GPT-Image-2Judge=MiMo-V2.52025.11 | 8.23 | 0.62 | 7.37 | 0.9 | 8.47 | 0.52 | 7.95 | 1.03 | 8.01 | 1 | |
| Nano-Banana-2Judge=MiMo-V2.52025.11 | 8.13 | 0.76 | 7.27 | 1.06 | 8.51 | 0.5 | 7.63 | 1.26 | 7.89 | 4 | |
| Nano-BananaJudge=MiMo-V2.52025.11 | 8.07 | 0.69 | 7.12 | 1.03 | 8.52 | 0.5 | 7.82 | 1.14 | 7.88 | 5 | |
| Med-Banana (Qwen)Judge=MiMo-V2.52025.11 | 8.07 | 0.7 | 7.06 | 1.07 | 8.57 | 0.5 | 8.02 | 1.02 | 7.93 | 2 | |
| Med-Banana (FireRed)Judge=MiMo-V2.52025.11 | 8.03 | 0.67 | 7.03 | 1.02 | 8.56 | 0.52 | 8.11 | 0.93 | 7.93 | 2 | |
| Qwen-Image-Edit-2512Judge=MiMo-V2.52025.11 | 7.95 | 0.65 | 7 | 1.01 | 8.46 | 0.5 | 7.65 | 1.14 | 7.77 | 6 | |
| FireRed-Image-Edit-1.1Judge=MiMo-V2.52025.11 | 7.91 | 0.7 | 7.06 | 1.02 | 8.36 | 0.48 | 7.44 | 1.04 | 7.69 | 7 | |
| GPT-Image-2Judge=Expert2025.11 | 7 | 2.17 | 6.5 | 2.03 | 7.07 | 1.79 | 7.07 | 2.02 | 6.91 | 1 | |
| Nano-Banana-2Judge=Expert2025.11 | 6.72 | 2.74 | 6.44 | 2.54 | 6.61 | 2 | 7 | 2.11 | 6.69 | 2 | |
| Med-Banana (Qwen)Judge=Expert2025.11 | 6.31 | 1.36 | 5.31 | 1.31 | 5.06 | 1.68 | 4.69 | 1.83 | 5.34 | 4 | |
| Med-Banana (FireRed)Judge=Expert2025.11 | 5.84 | 1.69 | 5.26 | 1.55 | 5.26 | 1.48 | 5.05 | 2.06 | 5.36 | 3 | |
| GPT-Image-2Judge=MedGemma-27B2025.11 | 5.76 | 2.1 | 3.84 | 1.73 | 6.47 | 2.32 | 3.84 | 1.5 | 4.98 | 1 | |
| Nano-Banana-2Judge=MedGemma-27B2025.11 | 5.75 | 2.17 | 3.92 | 1.94 | 5.9 | 2.81 | 3.71 | 1.82 | 4.82 | 2 | |
| Nano-BananaJudge=Expert2025.11 | 5.63 | 1.74 | 4.44 | 2.54 | 5.61 | 1 | 4 | 2.11 | 4.92 | 6 | |
| Qwen-Image-Edit-2512Judge=Expert2025.11 | 5.33 | 1.94 | 4.89 | 1.76 | 5.22 | 2.1 | 5.11 | 2.45 | 5.14 | 5 | |
| FireRed-Image-Edit-1.1Judge=Expert2025.11 | 5.2 | 1.94 | 4.47 | 1.75 | 4.4 | 1.58 | 4 | 1.93 | 4.52 | 7 | |
| Nano-BananaJudge=MedGemma-27B2025.11 | 4.57 | 1.72 | 2.81 | 1.09 | 5.52 | 2.87 | 3.09 | 1.84 | 4 | 4 | |
| Med-Banana (FireRed)Judge=MedGemma-27B2025.11 | 4.47 | 1.52 | 2.73 | 1.01 | 5.86 | 2.56 | 2.84 | 1.28 | 3.97 | 5 | |
| Med-Banana (Qwen)Judge=MedGemma-27B2025.11 | 4.27 | 1.72 | 2.72 | 1.13 | 6.19 | 2.58 | 3.21 | 1.85 | 4.1 | 3 | |
| Qwen-Image-Edit-2512Judge=MedGemma-27B2025.11 | 3.28 | 1.8 | 1.97 | 1.11 | 4.71 | 3.2 | 2.19 | 1.78 | 3.04 | 6 | |
| FireRed-Image-Edit-1.1Judge=MedGemma-27B2025.11 | 2.97 | 1.85 | 1.78 | 1.18 | 3.72 | 3.19 | 2.34 | 2.4 | 2.7 | 7 |