What is MiniCPM-OCR?
MiniCPM-OCR is a family of models for efficient document parsing and text recognition developed by OpenBMB OpenBMB/MiniCPM-OCR. 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. It works with different parsing models to help users locate content that needs review.
What features does MiniCPM-OCR 1.0 offer?
MiniCPM-OCR 1.0 (1.3B) is the first OCR-specific model in the MiniCPM-V/o family. It focuses on text-centric document understanding, covering text, tables, formulas, charts, seals, and structured information extraction.
Among open-source OCR models, MiniCPM-OCR 1.0 achieves 1.2 pages/s and the highest overall performance on OmniDocBench v1.6, Real5-OmniDocBench, Wild-OmniDocBench, and olmOCR-Bench. With DSpark decoding acceleration, it improves end-to-end processing efficiency by 55%.
How does MiniCPM-OCR Jev work?
MiniCPM-OCR Jev is an OCR correctness decision model for text, formulas, and tables. Given an image and its corresponding OCR result, it returns a quality score from 0 to 1 to help select reliable results and locate content that needs review.
With only 2B parameters, MiniCPM-OCR Jev achieves an AUROC of 93.6 on JevOCRBench, outperforming larger models and maintaining consistent generalization across multiple OCR engines.
How do I install and use MiniCPM-OCR?
You can integrate MiniCPM-OCR 1.0 and Jev models using various backend engines and development tools. Follow these steps to set up the environment and run your first document parsing task.
- Clone the repository or use the provided coding agent skill instructions.
- Choose an inference backend such as vLLM, SGLang, llama.cpp, or Ollama.
- Install the pip SDK for application integration and customization.
- Run local deployment or connect to the free API for testing.