Multimodal Understanding on MME (sum)
2,494.5Sum ScoreInternVL2.5-78B
Evaluation Results
| Method | Links | |
|---|---|---|
| InternVL2.5-78BModel Name=InternVL2.5-78B2024.12 | 2,494.5 | |
| Qwen2-VL-72BModel Name=Qwen2-VL-72B2024.12 | 2,482.7 | |
| InternVL2.5-38BModel Name=InternVL2.5-38B2024.12 | 2,455.8 | |
| InternVL2-Llama3-76BModel Name=InternVL2-Llama3-76B2024.12 | 2,414.7 | |
| InternVL2.5-26BModel Name=InternVL2.5-26B2024.12 | 2,373.3 | |
| MiniCPM-V2.6Model Name=MiniCPM-V2.62024.12 | 2,348.4 | |
| VanillaMode=Upper bound2024.12 | 2,347 | |
| InternVL2.5-8BModel Name=InternVL2.5-8B2024.12 | 2,344.1 | |
| InternVL2.5-4BModel Name=InternVL2.5-4B2024.12 | 2,337.5 | |
| iLLaVAImage token reduction ratio=66.7%2024.12 | 2,328 | |
| Qwen2-VL-7BModel Name=Qwen2-VL-7B2024.12 | 2,326.8 | |
| InternVL2-40BModel Name=InternVL2-40B2024.12 | 2,307.5 | |
| SparseVLMImage token reduction ratio=66.7%2024.12 | 2,299 | |
| VisionZipImage token reduction ratio=66.7%2024.12 | 2,299 | |
| VisionZipImage token reduction ratio=77.8%2024.12 | 2,280 | |
| PyramidDropImage token reduction ratio=66.7%2024.12 | 2,267 | |
| iLLaVAImage token reduction ratio=77.8%2024.12 | 2,266 | |
| LLaVA-OneVision-72BModel Name=LLaVA-OneVision-72B2024.12 | 2,261 | |
| InternVL2-26BModel Name=InternVL2-26B2024.12 | 2,260.7 | |
| FasterVLMImage token reduction ratio=66.7%2024.12 | 2,258 | |
| InternVL2-8BModel Name=InternVL2-8B2024.12 | 2,210.3 | |
| FasterVLMImage token reduction ratio=77.8%2024.12 | 2,210 | |
| SparseVLMImage token reduction ratio=77.8%2024.12 | 2,208 | |
| iLLaVAImage token reduction ratio=88.9%2024.12 | 2,207 | |
| InternVL-Chat-V1.5Model Name=InternVL-Chat-V1.52024.12 | 2,194.2 | |
| PyramidDropImage token reduction ratio=77.8%2024.12 | 2,178 | |
| VisionZipImage token reduction ratio=88.9%2024.12 | 2,159 | |
| InternVL2.5-2BModel Name=InternVL2.5-2B2024.12 | 2,138.2 | |
| FasterVLMImage token reduction ratio=88.9%2024.12 | 2,131 | |
| SparseVLMImage token reduction ratio=88.9%2024.12 | 2,128 | |
| InternVL2-4BModel Name=InternVL2-4B2024.12 | 2,059.8 | |
| PyramidDropImage token reduction ratio=88.9%2024.12 | 2,039 | |
| InternVL2.5-1BModel Name=InternVL2.5-1B2024.12 | 1,950.5 | |
| GPT-4VModel Name=GPT-4V2024.12 | 1,926.6 | |
| InternVL2-2BModel Name=InternVL2-2B2024.12 | 1,876.8 | |
| Qwen2-VL-2BModel Name=Qwen2-VL-2B2024.12 | 1,872 | |
| InternVL2-1BModel Name=InternVL2-1B2024.12 | 1,794.4 | |
| Claude-3-OpusModel Name=Claude-3-Opus2024.12 | 1,586.8 | |
| LLaVA-OneVision-0.5BModel Name=LLaVA-OneVision-0.5B2024.12 | 1,438 |