Image Understanding on SEED-Bench image
83.1AccuracyLumina-DiMOO
Evaluation Results
| Method | Links | |
|---|---|---|
| Lumina-DiMOOModel Architecture Style=Diff. Based2026.03 | 83.1 | |
| LFM2-VL-3B# Total Params=3.0B, Evaluation Toolkit=VLMEvalKit2025.11 | 76.55 | |
| InternVL3.5-2B# Total Params=2.3B, Evaluation Toolkit=VLMEvalKit2025.11 | 75.41 | |
| LLaDA-oModel Architecture Style=Diff. Based2026.03 | 75.3 | |
| LLaDA-VModel Architecture Style=Diff. Based2026.03 | 74.8 | |
| Qwen3-VL-2B# Total Params=2.1B, Evaluation Toolkit=VLMEvalKit2025.11 | 74.08 | |
| Qwen2.5-VL-3B# Total Params=3.75B, Evaluation Toolkit=VLMEvalKit2025.11 | 73.88 | |
| Janus-ProModel Architecture Style=AR Based2026.03 | 72.1 | |
| MetaMorphModel Architecture Style=AR Based2026.03 | 71.8 | |
| SmolVLM2-2.2B# Total Params=2.2B, Evaluation Toolkit=VLMEvalKit2025.11 | 71.3 | |
| Emu3Model Architecture Style=AR Based2026.03 | 68.2 | |
| MMaDAModel Architecture Style=Diff. Based2026.03 | 64.2 | |
| CASHEW-RLBackbone=Qwen3-VL-8B, Iterations (T)=32026.01 | 0.808 | |
| CASHEWBackbone=Qwen3-VL-8B2026.01 | 0.803 | |
| CASHEWBackbone=Qwen3-VL-4B2026.01 | 0.798 | |
| CASHEW-RLBackbone=Qwen3-VL-8B, Iterations (T)=12026.01 | 0.798 | |
| CASHEWBackbone=Qwen2.5-VL-7B2026.01 | 0.792 | |
| CASHEWBackbone=InternVL3.5-8B2026.01 | 0.79 | |
| Qwen3-VL-4BBackbone=Qwen3-VL-4B2026.01 | 0.787 | |
| Qwen3-VL-8BBackbone=Qwen3-VL-8B2026.01 | 0.787 | |
| InternVL3.5-8BBackbone=InternVL3.5-8B2026.01 | 0.777 | |
| Qwen2.5-VL-7BBackbone=Qwen2.5-VL-7B2026.01 | 0.775 | |
| LLaVA-OneVision-7BBackbone=LLaVA-OneVision-7B2026.01 | 0.754 | |
| LLaVA-Next-7BBackbone=LLaVA-Next-7B2026.01 | 0.724 | |
| LLaVA-1.5-7BBackbone=LLaVA-1.5-7B2026.01 | 0.661 | |
| VILA1.5-13BBackbone=VILA1.5-13B2026.01 | 0.628 | |
| Qwen-VL-Chat-7BBackbone=Qwen-VL-Chat-7B2026.01 | 0.582 |