The MiniCPM-OCR 1.0 model is a new 1.3-billion-parameter AI for optical character recognition, extracting text from documents, tables, and charts in English…

​ BEIJING, Oct. 31, 2026 /PRNewswire/ -- ModelBest and OpenBMB jointly announced the release and open-source availability of the MiniCPM-OCR series, a family of models for efficient document parsing and parsing-quality assessment. The launch includes two models: MiniCPM-OCR 1.0 (1.3B) and MiniCPM-OCR Jev (2B). As the first OCR-focused model in the MiniCPM family, MiniCPM-OCR 1.0 brings top-tier parsing accuracy to a compact 1.3-billion-parameter architecture optimized for edge deployment, pushing forward the balance between efficiency and performance for open-source OCR models.

Fast, accurate OCR with correctness assessment. Contribute to OpenBMB/MiniCPM-OCR development by creating an account on GitHub.

BEIJING, Oct. 31, 2026 /PRNewswire/ — ModelBest and OpenBMB jointly announced the release and open-source availability of the MiniCPM-OCR series, a family of models for efficient document parsing and parsing-quality assessment. The launch includes two models: MiniCPM-OCR 1.0 (1.3B) and MiniCPM-OCR Jev (2B). As the first OCR-focused model in the MiniCPM family, MiniCPM-OCR 1.0 brings […]

ModelBest and OpenBMB unveil the MiniCPM-OCR series, offering advanced document parsing solutions with AI-driven quality control for businesses.

Fast parsing. High accuracy.

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

OpenBMB has published model pages for MiniCPM-OCR-1.0 and MiniCPM-OCR-1.0-DSpark on Hugging Face, ModelScope and Modelers. The Hugging Face listing for the DSpark variant gives its task as image-text-to-text.

A image text to text model from openbmb with 190 likes and 791 downloads on Hugging Face.

Your OCR says it’s done—but did it get everything right? 🤔 In high-stakes workflows, even a small error can be costly. 🚀 Meet the MiniCPM-OCR suite: an open-source solution to parse fast and gu…

A 2-billion-parameter model that doesn't read documents, it judges whether another model's reading is right. OpenBMB releases MiniCPM-OCR Jev, scoring 0 to 1, with open weights on Hugging Face. Useful

YouTube video by Zero to MVP

Modelers.cn, the Huawei-backed platform powering the Ascend AI ecosystem, has just introduced a promising new asset for developers: DevQuasar/openbmb.MiniCPM-V-4.6-GGUF. This release packages…

深入探索前沿的MiniCPM-V系列模型，通过这门课程，掌握与MiniCPM-Llama3-V 2.5和MiniCPM-V 2.0模型相关的先进技术。本课程专注于模型的部署、优化及实际应用，适合进阶AI从业者。 课程亮点： MiniCPM-V模型简介 ：全面了解MiniCPM-Llama3-V 2.5和MiniCPM-V 2.0模型的基本概念、参数规模、性能指标，以及它们在OCR和多语言对话系统中的应用。掌握这些多模态大型语言模型（MLLMs）的技术优势及实际应用效果。 基于Gradio和Streamlit的Web端问答交互 ：掌握如何使用MiniCPM-Llama3-V2.5模型实现实时Web端问答系统。本模块涵盖Gradio和Streamlit环境的搭建与配置、交互界面的创建以及模型推理操作，提升用户交互体验。 基于llama.cpp的模型量化与推理 ：探讨使用llama.cpp框架对MiniCPM-Llama3-V2.5模型进行量化的技术。学习8位和4位量化方法，以及量化后模型在资源受限环境中的性能提升。 基于vLLM的模型推理与源码分析 ：深入了解MiniCPM-V模型在vLLM框架下的推理过程。解析模型的核心源代码，掌握有效模型推理的技术细节。 MiniCPM-V的全参数微调与LoRA微调 ：获得MiniCPM-V模型微调的实践经验。本节课涵盖全参数微调和LoRA技术，使用transformers Trainer和DeepSpeed配置微调脚本，自定义训练数据集，并应用高级微调方法以实现最佳性能。 通过本课程，你将全面掌握MiniCPM-V模型的部署与优化技巧，具备Web端互动、模型量化及高级微调技术的实战经验。 Who this course is for: AI技术爱好者：对多模态学习和最新AI技术有浓厚兴趣，希望深入了解和实践MiniCPM-V模型的技术细节。 数据科学家与机器学习工程师：希望拓展技能，掌握如何在实际项目中应用多模态模型，提高数据处理和模型融合的能力。 研究人员：从事AI和计算机视觉、自然语言处理领域的研究，寻找实用的多模态模型解决方案，提升科研能力。 学生与学习者：计算机科学、人工智能或相关领域的学生，想要通过实际项目获得实践经验，增强理论知识的应用能力。