Object Resize on curated synthetic benchmark (test)
0.3501Scale ErrorGPT-Image-1
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| GPT-Image-12026.01 | 0.3501 | 5.08 | 1.39 | |
| TALK2MOVE2026.01 | 0.3894 | 9.17 | 63.89 | |
| QwenImageEdit2026.01 | 0.4203 | 7.5 | 11.11 | |
| Flux-Kontext2026.01 | 0.4318 | 1.67 | 15.28 | |
| Bagel2026.01 | 0.479 | 0 | 8.33 |