| Benchmark | Model | Rank | Results |
|---|---|---|---|
| visual-question-answering-vqa-on-seed-seedbench | LatentUM (Base) | #12 | Accuracy: 76.3 |
| multimodal-reasoning-on-mmmu | LatentUM (Base) | #78 | Accuracy: 54.6 |
| text-to-image-generation-on-geneval | LatentUM Base | – | Overall: 0.85 |
| text-to-image-generation-on-geneval | LatentUM Vis-Gen | – | Overall: 0.87 |