OCRRouter

Mistral OCR vs dots.ocr

Pick Mistral OCR for high-volume PDF to Markdown through a hosted API, at $2 per 1,000 pages or $1 with the Batch API. Pick dots.ocr if you need to self-host an MIT-licensed 3B model that returns layout boxes, categories, text, tables as HTML and formulas as LaTeX in one pass. The main split is hosted per-page pricing against running your own GPU.

Mistral OCRmistral/mistral-ocrdots.ocrrednote/dots-ocr
TypeOCR APIOpen source
Price25 credits / page2 credits / page
Per 1,000 pages$25.00$2.00
Typical speed0.9 s / page1.5 s / page
Best forHigh-volume PDF to MarkdownMultilingual document parsing
InputsPDF, imagesPDF, images
Languages100+ languages100+ languages
Max pages per request1,000500
Tables✓ Yes✓ Yes
HandwritingNoNo
Math✓ Yes✓ Yes
Bounding boxes✓ Yes✓ Yes
ConfidenceNoNo
FormsNoNo
Multilingual✓ Yes✓ Yes
On OCRRouterAvailableAvailable

Prices are what you pay on OCRRouter. Speed is the provider's typical figure, not a benchmark we ran.

Which to choose

Choose Mistral OCR if

  • You want no infrastructure: it is a hosted API, so there is no GPU to run or maintain.
  • You process scanned business documents in bulk, and the Batch API halves OCR to $1 per 1,000 pages.
  • You need tables rebuilt as HTML with rowspan and colspan, which OCR 3 does.

Choose dots.ocr if

  • You need to keep documents on your own hardware: it is self-hosted through vLLM or Hugging Face Transformers, with MIT-licensed code.
  • You want layout bounding boxes, categories and text from one model, with formulas as LaTeX and tables as HTML.
  • You need chart or diagram conversion to SVG, which the dots.mocr-svg variant provides.

Key differences

Mistral OCRdots.ocr
How you run itHosted Mistral API (pinned version mistral-ocr-2512, OCR 3)Self-hosted via vLLM or Hugging Face Transformers; single CUDA GPU with vLLM is the documented path
LicenceProprietaryMIT, with a separate dots.ocr license agreement file in the repository
Price$2 per 1,000 pages; $3 with structured annotations; $1 per 1,000 with the Batch APIFree software; you pay for compute
Output formatsMarkdown, HTML, JSON, imagesJSON, Markdown, HTML, LaTeX, bounding boxes, SVG
BenchmarksOmniDocBench v1.5 overall 78.83 for the first version (independent); v1.6 overall 85.66olmOCR-bench 79.1 for dots.ocr and 83.9 for dots.mocr (reported by the authors)
Known failureHallucinated text on unreadable handwriting; reported misread characters and changed checkbox states on a dense financial tableOutput can repeat endlessly on runs of ellipses or underscores; a table above 1000 grid cells raises an error and drops the page in Docling
Model sizeNot published3B parameters (1.7B LLM backbone); no VRAM figure stated

Sources

What each one is

Mistral OCR

Fast document OCR that returns clean Markdown with tables and embedded images.

Mistral OCR details

dots.ocr

Single vision model from rednote that does layout detection and content recognition together.

dots.ocr details

Try both on your own file

Run the file once with each model and compare the output.

PDF, PNG, JPG, WebP, HEIC. Up to 20 MB each, 10 files.No file?

Use one, fall back to the other

{"model": "mistral/mistral-ocr","fallbacks": ["rednote/dots-ocr"],"file_url": "https://example.com/document.pdf"}

Questions

More comparisons

Mistral OCR alternativesdots.ocr alternativesBest OCR models in 2026