Efficient MLLMs | Tianyu Yu

The first complete OCR solution combining fast, accurate parsing with reliable OCR quality assessment.

MiniCPM-OCR 1.0 parses Chinese and English documents and PDFs, balancing quality and efficiency across challenging inputs such as scans, photographs, curved pages, and recaptured documents. MiniCPM-OCR Jev evaluates OCR results at the region level, distinguishing problematic regions from reliable ones.

MiniCPM-V 4.7

The first open-source MoE to run smoothly on-device without compromising performance. A new performance–efficiency Pareto frontier.

MiniCPM-V 4.6

Ultra-Efficient Architecture. MiniCPM-V 4.6 reduces the visual encoding computation FLOPs by more than 50%. It also supports mixed 4x/16x visual token compression rate, allowing flexible switching between accuracy and speed.

MiniCPM-o 4.5: Towards Real-Time Full-Duplex Omni-Modal Interaction

Real-time full-duplex omni-modal interaction via a unified streaming framework.

Junbo Cui, Bokai Xu, Chongyi Wang, Tianyu Yu, Weiyue Sun, Yingjing Xu, +28 more authors, Zhiyuan Liu, Yuan Yao Preprint

MiniCPM-o 4.5

The first full-duplex omni-modal large language model. It can continuously receive video and audio while generating speech and text, enabling real-time interactions that see, hear, and speak simultaneously.

MiniCPM-V 4.5: Cooking Efficient MLLMs via Architecture, Data, and Training Recipe

Efficient MLLM recipe across architecture, data, and training.

Tianyu Yu, Zefan Wang, Chongyi Wang, Fuwei Huang, Wenshuo Ma, Zhihui He, +26 more authors, Zhiyuan Liu, Maosong Sun CVPR 2026

MiniCPM-V 4.5

Efficient High-FPS and Long Video Understanding. Powered by a new unified 3D-Resampler over images and videos, MiniCPM-V 4.5 can now achieve 96x compression rate for video tokens, where 6 448x448 video frames can be jointly compressed into 64 video tokens (normally 1,536 tokens for most MLLMs).

Efficient GPT-4V Level Multimodal Large Language Model for Deployment on Edge Devices

Efficient GPT-4V-level MLLMs for edge-device deployment.

Yuan Yao, Tianyu Yu, Ao Zhang, Chongyi Wang, Junbo Cui, Hongji Zhu, +16 more authors, Zhiyuan Liu, Maosong Sun Nature Communications 2025