Mistral OCR 4.1

	Mistral OCR 4.1(docs.mistral.ai)
	403 points by spelk 8 days ago \| 160 comments
	tl;dr: Mistral released OCR 4.1 as a public preview, part of its Document AI stack. The update adds native paragraph-level bounding box extraction, structural block labels, and block-level confidence scores.
	HN Discussion: ↓Model underperforms compared to OpenAI pro models on complex documents like books with special typography ~Mistral OCR is not SOTA but excels at simple docs due to speed and affordability ↓Pricing is too expensive compared to alternatives like tesseract or self-hosted GPU pipelines ↓VLMs and OCR models have trust issues with censorship and hallucination on sensitive docs •Questions about multilingual support and comparison to alternatives like Baidu OCR