Image Editing Quality Assessment on AROURA-Bench
72.22AccuracyReasonEdit
Evaluation Results
| Method | Links | |
|---|---|---|
| ReasonEditZero-shot=true2026.05 | 72.22 | |
| LMM4EditZero-shot=true2026.05 | 65.21 | |
| EditReward (MiMo)Zero-shot=true2026.05 | 63.62 | |
| GPT-4oZero-shot=true2026.05 | 50.81 | |
| Gemini-2.5-FlashZero-shot=true2026.05 | 47.63 | |
| GPT-5Zero-shot=true2026.05 | 47.27 | |
| Gemini-2.0-FlashZero-shot=true2026.05 | 44.31 | |
| EditScore (Qwen3)Zero-shot=true2026.05 | 42.18 | |
| InternVL3.5 (7B)Zero-shot=true2026.05 | 36.92 | |
| Qwen3-VL (7B)Zero-shot=true2026.05 | 34.28 |