Mistral launches OCR 4, turning document extraction into a full enterprise AI play

Mistral AI released OCR 4, a document intelligence model that extracts not just text but structured document layouts with bounding boxes, block classifications, and confidence scores. The model supports 170 languages, handles formats including PDF, DOC, PPT, and OpenDocument, and can run in a single container on customer infrastructure for regulated enterprises.

OCR 4 reflects a shift from legacy optical character recognition toward document intelligence that preserves context, layout, and traceability for downstream AI systems. Its on-prem deployment and European positioning could appeal to banks, governments, and healthcare firms seeking AI tools that meet stricter data sovereignty and compliance requirements.