Multimodal Understanding on MM-Bench (MMB) en (dev)
85AccuracyBAGEL
Evaluation Results
| Method | Links | |
|---|---|---|
| BAGELModel Architecture Style=AR Based2026.03 | 85 | |
| Lumina-DiMOOModel Architecture Style=Diff. Based2026.03 | 84.5 | |
| LLaDA-VModel Architecture Style=Diff. Based2026.03 | 82.9 | |
| Show-o2Model Architecture Style=AR Based2026.03 | 79.3 | |
| Janus-ProModel Architecture Style=AR Based2026.03 | 79.2 | |
| LaViDa-OModel Architecture Style=Diff. Based2026.03 | 76.4 | |
| MetaMorphModel Architecture Style=AR Based2026.03 | 75.2 | |
| DimpleModel Architecture Style=Diff. Based2026.03 | 74.6 | |
| LLaDA-oModel Architecture Style=Diff. Based2026.03 | 71.1 | |
| LaViDa-LModel Architecture Style=Diff. Based2026.03 | 70.5 | |
| MMaDAModel Architecture Style=Diff. Based2026.03 | 68.5 | |
| Emu3Model Architecture Style=AR Based2026.03 | 58.5 |