Vision-language understanding on MMBench EN v1.1
89AccuracyMiniCPM-o 4.5
Evaluation results
| Method | Links | |
|---|---|---|
| MiniCPM-o 4.5Inference mode=Thinking mode, Size=9B2026.04 | 89 | |
| Qwen3-OmniInference mode=Thinking mode, Size=30B-A3B2026.04 | 88.2 | |
| Gemini 2.5 FlashInference mode=Thinking mode2026.04 | 87.1 | |
| GPT-5Inference mode=Thinking mode2026.04 | 85.5 | |
| Qwen3-VLInference mode=Thinking mode, Size=8B2026.04 | 85.3 |