Subject-driven generation on Constructed multimodal reasoning benchmark 350 samples (test)
0.3208CLIP-T ScoreDLA
Evaluation Results
| Method | Links | |
|---|---|---|
| DLA2026.05 | 0.3208 | |
| Qwen-Image2026.05 | 0.3099 | |
| DreamO2026.05 | 0.2888 | |
| UNO2026.05 | 0.2851 |
Loading the SOTA2 catalog…