| Benchmark | Model | Rank | Results |
|---|---|---|---|
| multimodal-reasoning-on-mmmu | VQRAE (InternViT-300M, Qwen2.5-7B, 448px) | – | Accuracy: 61.6 |
| text-to-image-generation-on-dpg-bench | VQRAE | – | DPG-Bench Score: 86.67 |
| text-to-image-generation-on-geneval | VQRAE | – | Overall: 0.76 |
| visual-question-answering-vqa-on-mmbench-en | VQRAE (InternViT-300M, Qwen2.5-7B, 448px) | – | Accuracy: 85.1 |