Introducing Mistral OCR 4

Mistral OCR 4 is a document processing model that extracts text from images and documents while also identifying where text is located, what type of content it represents, and how confident the model is in its extraction. The model supports 170 languages and can be deployed on a single container for organizations that need to keep document data within their own infrastructure. According to the source, independent annotators preferred OCR 4 over competing systems, and it achieves lower costs and faster processing speeds compared to some alternatives. The structured output is designed to support downstream tasks like retrieval-augmented generation, document-based agent actions, and content indexing pipelines.