Mistral OCR

Mistral OCR is a powerful document understanding service that enables users to extract interleaved text and images from a wide range of document formats. With its large 32,768 token context window, this model can handle complex, multi-page documents with ease.

Supports Text, Image, Video, Audio, Transcription, and Text-to-Speech inputs and outputs. Specialized for document processing and analysis tasks.

Supports fine-tuning for custom applications, allowing users to adapt the model to their specific needs. Capable of generating structured output formats for seamless integration into downstream workflows.

Also available via the alias mistral-ocr-latest.