Mistral OCR

Mistral AI · 1 year ago

Mistral has released Mistral OCR, an optical character recognition API that extracts text, images, tables, and equations from documents and PDFs while supporting thousands of languages and scripts. The API costs 1,000 pages per dollar and processes up to 2,000 pages per minute, with benchmark scores reaching 94.89% overall accuracy compared to competitors' 83-90%. Organizations can now use the tool to convert document repositories into machine-readable formats for applications like scientific research digitization, customer service automation, and legal document processing.