Foundations

Mistral Releases OCR 4 for Document Intelligence and Retrieval Pipelines

Mistral launched OCR 4 with layout data, typed blocks and confidence scores for search, RAG and document-processing workflows.

Mistral introduced OCR 4, a document-intelligence model that returns extracted text alongside bounding boxes, block classifications and confidence scores. The company positioned it as an ingestion component for enterprise search, RAG and domain-specific retrieval.

The release illustrated why document processing remains a core AI infrastructure problem. An agent cannot reliably reason over an enterprise record if the system first loses layout, tables or confidence information while turning a document into text.

Editorial sources

Every claim in this briefing traces back to the references below.

More from the brief

Browse all →
Guides

88 Hours, 130 Billion Tokens: Inside OpenAI's Navier-Stokes Proof

Guides

$900M ARR and a $48B Bet: What Cognition's Round Says About Agent Demand

Guides

Microsoft Puts Day-Long Agent Work Inside a Windows 365 Cloud PC