Multimodal Multilingual Reasoning on MMMB
86.8English AccuracyQwen2-VL-72B
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Qwen2-VL-72BParameters=72B2024.12 | 86.8 | 85.3 | 85.2 | 84.8 | 84.2 | 85.3 | — | |
| InternVL2.5-38BParameters=38B2024.12 | 86.4 | 85.1 | 84.1 | 84.3 | 82.8 | 84.9 | — | |
| InternVL2.5-78BParameters=78B2024.12 | 86.3 | 85.6 | 85.1 | 84.8 | 83.1 | 85.4 | — | |
| InternVL2.5-26BParameters=26B2024.12 | 86.2 | 83.8 | 81.6 | 73.3 | 73.7 | 82.8 | — | |
| LaV-CoT-3Btraining=SFT + GRPO2025.09 | 86 | 84.8 | 83 | 82.7 | 80.3 | 83.6 | — | |
| InternVL2-40BParameters=40B2024.12 | 85.3 | 84.1 | 81.1 | 70.3 | 74.2 | 81.4 | — | |
| InternVL2-Llama3-76BBackbone=Llama3, Parameters=76B2024.12 | 85.3 | 85.1 | 82.8 | 82.8 | 83 | 83.7 | — | |
| Gemini-2.5-flash2025.09 | 85.2 | 84.5 | 83.1 | 81.8 | 79.6 | 84 | — | |
| InternVL3-8B2025.09 | 85.1 | 83.1 | 82.5 | 81.6 | 76.2 | 83.4 | — | |
| Qwen2.5-VL-7B2025.09 | 85 | 83.6 | 82.1 | 83.3 | 76.4 | 83.2 | — | |
| InternVL3.5-8B2025.09 | 84.9 | 83 | 81.4 | 79.6 | 77.4 | 83.1 | — | |
| GPT-4o-2024-05-132025.09 | 84.9 | 84.3 | 82.8 | 82.3 | 79 | 83.3 | — | |
| Eagle2-9B2025.09 | 84.4 | 83 | 81.8 | 81.2 | 75.8 | 83 | — | |
| InternVL2.5-8BParameters=8B2024.12 | 84.3 | 83.1 | 78.6 | 69.3 | 71.5 | 79.5 | — | |
| Qwen2-VL-7BParameters=7B2024.12 | 83.9 | 82.4 | 81.2 | 79 | 74.7 | 82.4 | — | |
| Qwen2-VL-7B2025.09 | 83.9 | 82.4 | 81.2 | 79 | 74.7 | 82.4 | — | |
| InternVL2-26BParameters=26B2024.12 | 83.8 | 81.7 | 78 | 68.8 | 69.3 | 76.3 | — | |
| InternVL2.5-4BParameters=4B2024.12 | 83.7 | 81 | 79.7 | 76 | 70.5 | 79.9 | — | |
| InternVL2-8BParameters=8B2024.12 | 83.4 | 81.5 | 76.1 | 66.3 | 69.2 | 75.7 | — | |
| LaV-CoT-3Btraining=SFT2025.09 | 83.2 | 81.1 | 80.6 | 80.5 | 77.3 | 82.2 | — | |
| InternVL-Chat-V1.5Version=V1.52024.12 | 82.6 | 80.8 | 76.3 | 65.2 | 68.6 | 74 | — | |
| InternVL2-4BParameters=4B2024.12 | 82 | 76.1 | 75.6 | 54.3 | 51.2 | 67.4 | — | |
| InternVL3-2B2025.09 | 81.9 | 78.3 | 75.4 | 68.6 | 62.9 | 74.6 | — | |
| InternVL2.5-2BParameters=2B2024.12 | 81.4 | 74.4 | 58.2 | 48.3 | 46.4 | 53.2 | — | |
| Qwen2.5-VL-3B2025.09 | 81.2 | 81 | 74.8 | 71 | 65.6 | 75.2 | — | |
| InternVL3.5-2B2025.09 | 80.2 | 77.7 | 75.9 | 68.5 | 69.1 | 76.3 | — | |
| PARROT-7B2025.09 | 80.1 | 80 | 79.6 | 76.5 | 75 | 79.9 | — | |
| InternVL2-2BParameters=2B2024.12 | 79.4 | 71.6 | 54 | 43.5 | 46.4 | 48.1 | — | |
| LLaVA-OneVision-7B2025.09 | 79 | 78.2 | 75.9 | 73.3 | 67.7 | 76.3 | — | |
| InternVL2.5-1BParameters=1B2024.12 | 78.8 | 70.2 | 61.5 | 55 | 45.3 | 61.1 | — | |
| Qwen2-VL-2BParameters=2B2024.12 | 78.3 | 74.2 | 72.6 | 68.3 | 61.8 | 72.8 | — | |
| Qwen2-VL-2B2025.09 | 78.3 | 74.2 | 72.6 | 68.3 | 61.8 | 72.8 | — | |
| GPT-4V2024.12 | 75 | 74.2 | 71.5 | 73.5 | 69 | 73.1 | — | |
| Gemini-1.0-Pro2024.12 | 75 | 71.9 | 70.6 | 69.9 | 69.6 | 72.7 | — | |
| InternVL2-1BParameters=1B2024.12 | 73.2 | 67.4 | 55.5 | 53.5 | 43.8 | 55.2 | — | |
| DeepSeek-VL-7B2025.09 | 72.6 | 65.9 | 64.4 | 49.7 | 49 | 67.5 | — | |
| GLM-4v-9B2025.09 | 69.2 | 62.8 | 61.5 | 47.2 | 46.9 | 64.3 | — | |
| mPLUG-Owl22024.12 | 67.3 | 61 | 59.7 | 45.8 | 45.4 | 62.6 | — | |
| Monkey-9.8B2025.09 | 66 | 58.1 | 46.3 | 38.8 | 37.6 | 48.5 | — |