Multimodal Perception and Cognition on MME (test)
2,396Overall ScoreMulberry-7B
Evaluation results
| Method | Links | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mulberry-7B2026.04 | 2,396 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen2.5-VL-7B + Saliency-R1Training Stage=Saliency-R12026.04 | 2,385 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen2.5-VL-7B + SFTTraining Stage=SFT2026.04 | 2,372 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen2.5-VL-3B + SFTTraining Stage=SFT2026.04 | 2,351 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MiniCPM-V-2.6-8B2026.04 | 2,348 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| InternVL2.5-8B2026.04 | 2,344 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GPT-4o2026.04 | 2,329 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Insight-V-8B2026.04 | 2,312 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Upper BoundBackbone=Qwen2.5-VL-7B2026.04 | 2,308 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen2.5-VL-7BTraining Stage=Base2026.04 | 2,302 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen2.5-VL-3B + Saliency-R1Training Stage=Saliency-R12026.04 | 2,235 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| IXC-2.52026.04 | 2,229 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| InternVL2-8B2026.04 | 2,210 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Qwen2.5-VL-3BTraining Stage=Base2026.04 | 2,206 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CLASPBackbone=Qwen2.5-VL-7B, Token Pruning Rate=66.7%2026.04 | 2,197 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Vision-R1-LlamaV-CI-11B2026.04 | 2,190 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SparseVLMBackbone=Qwen2.5-VL-7B, Token Pruning Rate=66.7%2026.04 | 2,165 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CLASPBackbone=Qwen2.5-VL-7B, Token Pruning Rate=77.8%2026.04 | 2,154 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SparseVLMBackbone=Qwen2.5-VL-7B, Token Pruning Rate=77.8%2026.04 | 2,086 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MiniCPM-Llama-V-2.5-8B2026.04 | 2,025 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CLASPBackbone=Qwen2.5-VL-7B, Token Pruning Rate=88.9%2026.04 | 2,010 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LLaVA-OneVision-7B2026.04 | 1,998 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Claude-3.5 Sonnet2026.04 | 1,920 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| InternLM-XComposerModel size=7B2023.09 | 1,919.5 | 190 | 158.3 | 126.7 | 165 | 125 | 161.9 | 150.3 | 159.8 | 165.3 | 126.3 | 138.6 | 55 | 112.5 | 85 | 137.1 | |
| Qwen-VL-ChatModel size=7B2023.09 | 1,848.3 | 158.3 | 150 | 128.3 | 170 | 140 | 178.6 | 120.6 | 152.3 | 164 | 125.5 | 130.7 | 40 | 147.5 | 42.5 | 132 | |
| LLaVA-1.5Model size=7B2023.09 | 1,826.7 | 185 | 155 | 133.3 | 170 | 125 | 160.5 | 152.9 | 161.3 | 170.5 | 117.7 | 127.8 | 42.5 | 77.5 | 47.5 | 130.5 | |
| SparseVLMBackbone=Qwen2.5-VL-7B, Token Pruning Rate=88.9%2026.04 | 1,813 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MMICLModel size=7B2023.09 | 1,810.7 | 170 | 160 | 81.7 | 156.7 | 100 | 146.3 | 141.8 | 153.8 | 136.1 | 135.5 | 136.4 | 82.5 | 132.5 | 77.5 | 129.3 | |
| Cambrain-1-8B2026.04 | 1,803 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| BLIVAModel size=7B2023.09 | 1,669.2 | 180 | 138.3 | 81.7 | 180 | 87.5 | 155.1 | 140.9 | 151.5 | 89.5 | 133.3 | 136.4 | 57.5 | 77.5 | 60 | 119.2 | |
| CheetorModel size=7B2023.09 | 1,584.4 | 180 | 96.7 | 80 | 80 | 100 | 147.3 | 164.1 | 156 | 145.7 | 113.5 | 98.6 | 77.5 | 57.5 | 87.5 | 113.2 | |
| OtterModel size=7B2023.09 | 1,572 | 195 | 88.3 | 86.7 | 86.7 | 72.5 | 138.8 | 172.7 | 158.8 | 137.3 | 129 | 106.4 | 72.5 | 57.5 | 70 | 112.3 | |
| LynxModel size=7B2023.09 | 1,508.9 | 195 | 151.7 | 90 | 90 | 77.5 | 124.8 | 118.2 | 164.5 | 162 | 119.5 | 110.7 | 17.5 | 42.5 | 45 | 107.8 | |
| InstructBLIPModel size=7B2023.09 | 1,417.9 | 185 | 143.3 | 66.7 | 66.7 | 72.5 | 123.8 | 101.2 | 153 | 79.8 | 134.3 | 129.3 | 40 | 65 | 57.5 | 101.3 | |
| mPLUG-OwlModel size=7B2023.09 | 1,238.4 | 120 | 50 | 50 | 50 | 65 | 136.1 | 100.3 | 135.5 | 159.3 | 96.3 | 78.6 | 60 | 80 | 57.5 | 88.5 | |
| LLaMA-A.-V2Model size=7B2023.09 | 1,194.9 | 120 | 50 | 48.3 | 48.3 | 125 | 99.7 | 86.2 | 148.5 | 150.3 | 69.8 | 81.4 | 62.5 | 50 | 55 | 85.4 | |
| VisualGLMModel size=7B2023.09 | 880.4 | 85 | 50 | 48.3 | 48.3 | 42.5 | 66 | 53.2 | 146.3 | 83.8 | 75.3 | 39.3 | 45 | 50 | 47.5 | 62.9 | |
| LLaVAModel size=7B2023.09 | 712.5 | 50 | 50 | 50 | 50 | 50 | 50 | 48.8 | 50 | 50 | 49 | 57.1 | 50 | 57.5 | 50 | 50.9 | |
| MiniGPT-4Model size=7B2023.09 | 694.3 | 68.3 | 55 | 43.3 | 43.3 | 57.5 | 41.8 | 54.4 | 71.8 | 54 | 60.5 | 59.3 | 45 | 0 | 40 | 49.6 |