Multi-modal Understanding on MME (MME score)
2,371.9MME ScoreInternVL3.5-8B
Evaluation Results
| Method | Links | |
|---|---|---|
| InternVL3.5-8BRatio=100%, Backbone=InternVL3.5-8B2026.05 | 2,371.9 | |
| VanillaVisual Tokens=866, Prefill Time (s)=0.2215, FLOPs (T)=2.1015, GPU Mem. (GB)=16.01, SpeedUp=1.00×2026.05 | 2,329 | |
| DivPruneRatio=60%, Backbone=InternVL3.5-8B2026.05 | 2,298.3 | |
| F3ARatio=60%, Backbone=InternVL3.5-8B2026.05 | 2,291.7 | |
| BaselineArchitecture=Qwen-2.5-VL2026.04 | 2,279 | |
| CDPrunerRatio=60%, Backbone=InternVL3.5-8B2026.05 | 2,276.1 | |
| Latent DenoisingArchitecture=Qwen-2.5-VL2026.04 | 2,265 | |
| FastVRatio=60%, Backbone=InternVL3.5-8B2026.05 | 2,264.9 | |
| VisionZipRatio=60%, Backbone=InternVL3.5-8B2026.05 | 2,260 | |
| SPprunerVisual Tokens=192, Prefill Time (s)=0.0876, FLOPs (T)=0.5798, GPU Mem. (GB)=15.86, SpeedUp=2.53×2026.05 | 2,182 | |
| DARTVisual Tokens=192, Prefill Time (s)=0.1068, FLOPs (T)=0.5798, GPU Mem. (GB)=15.86, SpeedUp=2.07×2026.05 | 2,174 | |
| FastVRatio=40%, Backbone=InternVL3.5-8B2026.05 | 2,167.9 | |
| CDPrunerRatio=40%, Backbone=InternVL3.5-8B2026.05 | 2,164.9 | |
| F3ARatio=40%, Backbone=InternVL3.5-8B2026.05 | 2,135.7 | |
| VisionZipRatio=40%, Backbone=InternVL3.5-8B2026.05 | 2,131.4 | |
| DivPruneRatio=40%, Backbone=InternVL3.5-8B2026.05 | 2,122.4 | |
| F3ARatio=20%, Backbone=InternVL3.5-8B2026.05 | 2,115.7 | |
| CDPrunerRatio=20%, Backbone=InternVL3.5-8B2026.05 | 2,101.5 | |
| DivPruneRatio=20%, Backbone=InternVL3.5-8B2026.05 | 2,099.1 | |
| VisionZipRatio=20%, Backbone=InternVL3.5-8B2026.05 | 2,051.7 | |
| FastVRatio=20%, Backbone=InternVL3.5-8B2026.05 | 1,968.7 | |
| BaselineArchitecture=LLaVA+CLIP2026.04 | 1,879 | |
| Latent DenoisingArchitecture=LLaVA+CLIP2026.04 | 1,846 | |
| Latent DenoisingArchitecture=LLaVA+SigLIP2026.04 | 1,733 | |
| BaselineArchitecture=LLaVA+SigLIP2026.04 | 1,708 |