
Mistral AI has released OCR 4, a document-reading model that identifies text location and structural role within PDFs, Word, and PowerPoint files. In a blind test comparing it against competitors on over 600 documents, independent reviewers preferred OCR 4 72 percent of the time. The model supports 170 languages and is available now through multiple platforms at $4 per 1,000 pages ($2 in batch mode).
Summaries like this, in your inbox every morning.
Sign up free →What happened
Mistral AI launched OCR 4, a model that reads text from PDFs, Word files, and PowerPoint presentations while also identifying the structural role of each element (titles, tables, equations, signatures) and assigning confidence scores to its readings. In blind tests with over 600 documents, independent reviewers preferred OCR 4's output 72 percent of the time over competing models.
Why it matters
Unlike earlier OCR versions that extract raw text only, OCR 4 organizes documents into meaningful sections automatically—useful for feeding data into search systems or letting AI agents process them. The model supports 170 languages, including less common ones, making it potentially valuable for businesses that handle documents in multiple languages.
What to watch
OCR 4 is available now through the API, Mistral Studio, and Microsoft Foundry. Pricing is $4 per 1,000 pages in standard mode or $2 in batch mode.
No comments yet. Be the first to share your thoughts!
Log in to join the discussion





Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
5 minutes a day. The AI essentials.
200+ sources · Email / LINE / Slack