Multimodal Understanding on MMBench en (dev)
84.2ScoreQwen2.5-VL-7B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen2.5-VL-7BModel Category=Open-source Autoregressive Models2025.12 | 84.2 | — | |
| InternVL2_5-8BModel Category=Open-source Autoregressive Models2025.12 | 84.1 | — | |
| Qwen2.5VLSize=7B, Type=AR, Samples=>9M2025.12 | 83.5 | — | |
| DiffusionVLSize=7B, Type=Diff., Samples=738K2025.12 | 83.5 | — | |
| LLaDA-VSize=8B, Type=Diff., Samples=16.5M2025.12 | 82.9 | — | |
| LLaDA-V-8BModel Category=Discrete Diffusion Models, Controlled Group=true2025.12 | 82.9 | — | |
| InternVL2-8BModel Category=Open-source Autoregressive Models2025.12 | 82.7 | — | |
| InternVL2_5-4BModel Category=Open-source Autoregressive Models2025.12 | 82.6 | — | |
| SDAR-VL-8B-InstModel Category=Block-wise Discrete Diffusion Models2025.12 | 82.2 | — | |
| LLaVA-OV-7BModel Category=Open-source Autoregressive Models, Controlled Group=true2025.12 | 82.1 | — | |
| Qwen2-VL-7BModel Category=Open-source Autoregressive Models2025.12 | 81.7 | — | |
| InternVL-2-8BBackbone=InternVL-2-8B2026.01 | 81.7 | 602.8 | |
| InternVL-2-26BBackbone=InternVL-2-26B2026.01 | 81.6 | 600.1 | |
| LLaVA-OVSize=7B, Type=AR, Samples=7.8M2025.12 | 80.8 | — | |
| DiffusionVLSize=3B, Type=Diff., Samples=738K2025.12 | 80.1 | — | |
| Qwen2.5-VL-3BModel Category=Open-source Autoregressive Models2025.12 | 79.8 | — | |
| SLIBackbone=LLaVA-OV-7B2026.01 | 79.6 | 551.1 | |
| LLaVA-OV-7BBackbone=LLaVA-OV-7B2026.01 | 79.2 | 556.2 | |
| Qwen2.5VLSize=3B, Type=AR, Samples=>9M2025.12 | 79.1 | — | |
| CLIBackbone=LLaVA-OV-7B2026.01 | 78.9 | 559.4 | |
| SDAR-VL-4B-InstModel Category=Block-wise Discrete Diffusion Models2025.12 | 78.7 | — | |
| DeepStackBackbone=LLaVA-OV-7B2026.01 | 78.1 | 539.3 | |
| InternVL2-4BModel Category=Open-source Autoregressive Models2025.12 | 77.4 | — | |
| LaViDA-O-8BModel Category=Discrete Diffusion Models2025.12 | 76.4 | — | |
| Cambrian-1Size=8B, Type=AR, Samples=-2025.12 | 75.9 | — | |
| GPT-4VModel Category=Closed-source Models2025.12 | 75 | — | |
| Dimple-8BModel Category=Discrete Diffusion Models2025.12 | 74.6 | — | |
| LaViDa-LSize=8B, Type=Diff., Samples=1.6M2025.12 | 70.5 | — | |
| LaViDA-L-8BModel Category=Discrete Diffusion Models2025.12 | 70.5 | — | |
| MMaDA-8BModel Category=Discrete Diffusion Models2025.12 | 68.5 | — | |
| IXC-2.5-7BBackbone=IXC-2.5-7B2026.01 | 66.1 | 380.1 | |
| LLaVA-1.5Size=7B, Type=AR, Samples=-2025.12 | 64.3 | — | |
| VILA-13BBackbone=VILA-13B2026.01 | 57.9 | 396.2 | |
| DeepStackBackbone=LLaVA-OV-0.5B2026.01 | 51.2 | 380.9 | |
| CLIBackbone=LLaVA-OV-0.5B2026.01 | 48.3 | 395.6 | |
| SLIBackbone=LLaVA-OV-0.5B2026.01 | 47.3 | 393.7 | |
| LLaVA-OV-0.5BBackbone=LLaVA-OV-0.5B2026.01 | 45.8 | 391 | |
| LLaVASize=7B, Type=AR, Samples=-2025.12 | 38.7 | — |