LVLM Evaluation on MMStar
76.6CP ScoreGPT4V (high)
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| GPT4V (high)LLM=GPT4-Turbo, Param.=-2026.02 | 76.6 | 51.4 | 66.6 | 55.8 | 42.6 | 49.8 | 57.1 | — | |
| Qwen3VL-4B-ThinkingLLM=Qwen3-4B, Param.=4B2026.02 | 73.6 | 58.8 | 74.8 | 73.6 | 50.8 | 86.8 | 68.4 | — | |
| InternLM-XC2LLM=InternLM2-7B, Param.=7B2026.02 | 70.8 | 48.8 | 65.2 | 56.4 | 42 | 49.2 | 55.4 | — | |
| InternVL-Chat-V1.2LLM=NH2-Yi-34B, Param.=40B2026.02 | 67.6 | 43.2 | 61.2 | 47.2 | 24 | 19.2 | 43.7 | — | |
| CogVLM-ChatLLM=Vicuna-v1.5-7B, Param.=17B2026.02 | 66.8 | 36.8 | 49.2 | 31.2 | 23.6 | 11.6 | 36.5 | — | |
| LLaVA-NextLLM=NH2-Yi-34B, Param.=34B2026.02 | 66.4 | 52 | 62.4 | 46 | 32.4 | 53.6 | 52.1 | — | |
| Deepseek-VLLLM=Deepseek-7B, Param.=8B2026.02 | 64 | 30.8 | 49.2 | 36.4 | 21.6 | 20.4 | 37.1 | — | |
| GPT4V (low)LLM=GPT4-Turbo, Param.=-2026.02 | 62 | 32.8 | 55.2 | 48 | 33.6 | 44.8 | 46.1 | — | |
| TinyLLaVALLM=Phi2-2.7B, Param.=3B2026.02 | 60.4 | 31.6 | 50.8 | 30.4 | 18 | 24.8 | 36 | — | |
| Qwen-VL-ChatLLM=Qwen-7B, Param.=8B2026.02 | 59.6 | 32 | 50.8 | 29.2 | 22 | 31.6 | 37.5 | — | |
| LLaVA-1.5LLM=Vicuna-v1.5-7B, Param.=7B2026.02 | 58.8 | 24 | 38.8 | 24 | 13.6 | 22.8 | 30.3 | — | |
| ShareGPT4VLLM=Vicuna-v1.5-7B, Param.=7B2026.02 | 58.8 | 28 | 45.6 | 24.4 | 17.2 | 24 | 33 | — | |
| LLaVA-1.5LLM=Vicuna-v1.5-13B, Param.=13B2026.02 | 58.8 | 28 | 41.6 | 24.4 | 18.4 | 25.6 | 32.8 | — | |
| Sphinx-X-MOELLM=Mixtral-8x7B, Param.=57B2026.02 | 58.4 | 40.8 | 47.6 | 35.2 | 19.2 | 32 | 38.9 | — | |
| Yi-VLLLM=Yi-6B, Param.=6B2026.02 | 58 | 33.6 | 46.4 | 34.8 | 20.4 | 34 | 37.9 | — | |
| Monkey-ChatLLM=Qwen-7B, Param.=10B2026.02 | 57.6 | 36.4 | 51.6 | 33.2 | 26.4 | 24.4 | 38.3 | — | |
| Yi-VLLLM=Yi-34B, Param.=34B2026.02 | 53.2 | 31.2 | 52 | 32.4 | 12.4 | 35.2 | 36.1 | — | |
| GeminiPro-VisionLLM=GeminiPro, Param.=-2026.02 | 51.6 | 28.8 | 50.8 | 46 | 28.4 | 50 | 42.6 | — | |
| Random ChoiceLLM=-, Param.=-2026.02 | 23.7 | 24.5 | 25.3 | 24.3 | 24.8 | 25.1 | 24.6 | — | |
| ReMoT-CoT-4BLLM=Qwen3-4B, Param.=4B2026.02 | 3.2 | 60.8 | 2.8 | 78.4 | 55.2 | 2 | 70.4 | — | |
| Gemini 2.5 FlashSize=-, mode=instruct2026.04 | — | — | — | — | — | — | — | 75.8 | |
| InternVL3.5Size=8B, mode=instruct2026.04 | — | — | — | — | — | — | — | 69.3 | |
| MiniCPM-o 4.5Size=9B, mode=instruct2026.04 | — | — | — | — | — | — | — | 73.1 | |
| Qwen3-OmniSize=30B-A3B, mode=instruct2026.04 | — | — | — | — | — | — | — | 68.5 | |
| Qwen3-VLSize=8B, mode=instruct2026.04 | — | — | — | — | — | — | — | 70.9 |