Back to Gap Map Model components / Document conversion & OCR

Mistral OCR

closed / Overall score: n/a

Document-understanding API from Mistral that converts PDFs and images into ordered markdown with images, tables and equations preserved as structured elements, and returns them in a form intended to be fed straight into a model's context.

Openness

1 high confidence

1.0

license: Proprietary
service: proprietary(commercial managed service)
source: closed(no implementation is published)

It is a commercial managed service with no published implementation, and it cannot be self-hosted. There is no source at all to weigh here.

https://mistral.ai/news/mistral-ocr recorded 2026-09-16
Mistral's own announcement of Mistral OCR, describing a document-understanding API that extracts ordered markdown with interleaved images, tables and equations. Establishes a hosted commercial API with no published implementation.

Adoption

not assessed

No usage figure is published for this service specifically, and it publishes no countable artifact - no package, repository or registry entry - so no level is assigned rather than one being inferred from the vendor's platform as a whole.

https://mistral.ai/news/mistral-ocr recorded 2026-09-16
The product page, read for a usage disclosure; it publishes none specific to this service.

Capability

4 medium confidence

4.0

Mistral OCR returns ordered markdown with tables, equations and images preserved as structured elements, matching Chandra's capability. Nothing published lets a customer define their own entity types to extract, which is what would place it higher.

https://mistral.ai/news/mistral-ocr recorded 2026-09-16
Mistral's own announcement of Mistral OCR, describing a document-understanding API that extracts ordered markdown with interleaved images, tables and equations. Establishes a hosted commercial API with no published implementation.

Verified 2026-09-16