MiniCPM-V

<div align="center">

<a href="./docs/minicpm_v4dot7_en.md"><img src="./assets/minicpmv4.7/release-hero.svg" alt="MiniCPM-V 4.7 — A new chapter in on-device multimodal AI. 35B-A3B and 1B models for images, video and documents." width="100%"></a>

**MiniCPM-V & MiniCPM-o · See, hear, and interact on your local devices.**

<p>
  <strong><a href="./docs/minicpm_v4dot7_en.md">MiniCPM-V 4.7</a></strong>&nbsp;<a href="https://huggingface.co/openbmb/MiniCPM-V-4.7-35B-A3B" title="MiniCPM-V 4.7 35B-A3B on Hugging Face">🤗</a>&nbsp;35B-A3B&nbsp;<a href="https://huggingface.co/openbmb/MiniCPM-V-4.7-1B" title="MiniCPM-V 4.7 1B on Hugging Face">🤗</a>&nbsp;1B&nbsp;<a href="https://huggingface.co/spaces/openbmb/MiniCPM-V-4.7-Demo" title="MiniCPM-V 4.7 Demo">🤖</a>
  &emsp;
  <strong><a href="./docs/minicpm_o4dot5_en.md">MiniCPM-o 4.5</a></strong>&nbsp;<a href="https://huggingface.co/openbmb/MiniCPM-o-4_5" title="MiniCPM-o 4.5 on Hugging Face">🤗</a>&nbsp;<a href="https://minicpmo45.modelbest.cn" title="MiniCPM-o 4.5 Demo">🤖</a>
</p>

<p>
  <sub><a href="#model-overview">Overview</a> &nbsp;·&nbsp; <a href="#get-started">Get started</a> &nbsp;·&nbsp; <a href="#project-resources">Resources</a> &nbsp;|&nbsp; <a href="./README_zh.md">中文</a> · English</sub>
</p>

</div>

## News

- [2026.10.28] 🔥🔥🔥 We open-source the [MiniCPM-OCR suite](https://github.com/OpenBMB/MiniCPM-OCR), pairing two models for **fast, accurate parsing and result verification**: **MiniCPM-OCR 1.0** (based on **MiniCPM-V 4.6**, SOTA on Real5-OmniDocBench and OCRBench v2) and **MiniCPM-OCR Jev** for result correctness checks.
- **2026-10-15** · We release **DSpark speculative decoding heads** in **BF16 and GGUF** for both MiniCPM-V 4.7 variants: **1B** ([**DSpark-BF16**](https://huggingface.co/openbmb/MiniCPM-V-4.7-1B-DSpark), [**DSpark-GGUF**](https://huggingface.co/openbmb/MiniCPM-V-4.7-1B-DSpark-gguf)) and **35B-A3B** ([**DSpark-BF16**](https://huggingface.co/openbmb/MiniCPM-V-4.7-35B-A3B-DSpark), [**DSpark-GGUF**](https://huggingface.co/openbmb/MiniCPM-V-4.7-35B-A3B-DSpark-gguf)).
- **2026-10-12 · MiniCPM-V 4.7 released** — the first open-source MoE model to run smoothly on-device without compromising performance, setting a new multimodal performance–efficiency Pareto frontier. [See what's new](./docs/minicpm_v4dot7_en.md).

## MiniCPM-V 4.7

MiniCPM-V 4.7 is the latest model in the MiniCPM-V series, available in **35B-A3B and 1B variants**. The 35B-A3B variant is the series' first MoE model. Its main features include:

- **🚀 A new performance–efficiency Pareto frontier.** The 35B-A3B and 1B variants each advance the performance–efficiency Pareto frontier for multimodal large language models. In Instruct mode, 35B-A3B **achieves a higher overall score in less inference time than Qwen3.6 35B-A3B**; the 1B variant further improves on MiniCPM-V 4.6's core capabilities and **substantially outperforms Qwen3.5 0.8B in both overall score and inference efficiency**.
- **📱 The first open-source MoE to run