A Pocket-Sized MLLM for Ultra-Efficient Image and Video Understanding on Your Phone

Benchmark Model Rank Results
chart-question-answering-on-charxiv-rqMiniCPM-V 4.6-Thinking (16x)–Accuracy: 45.4
optical-character-recognition-ocr-on-ocrbenchMiniCPM-V 4.6-Thinking (4x)–Score: 857
optical-character-recognition-ocr-on-ocrbench-v2MiniCPM-V 4.6 (Instruct, 4x)–English Score: 40.6Chinese Score: 47.3
optical-character-recognition-ocr-on-omnidocbench-v1-5MiniCPM-V 4.6 (Instruct, 4x)–Accuracy: 84.6
visual-question-answering-vqa-on-blinkMiniCPM-V 4.6-Thinking (16x)–Accuracy: 58.7
visual-question-answering-vqa-on-mmbench-cnMiniCPM-V 4.6-Thinking (16x)–Accuracy: 81.0