Mistral OCR vs Pixtral
Pick Mistral OCR for any new OCR work, such as bulk PDF to Markdown for RAG or batch scanning of business documents. It is a dedicated OCR API with a $2 per 1,000 pages list price, and it is the model Mistral still maintains. Pixtral 12B is a general vision model that Mistral has deprecated, so keep it only for existing integrations or for question answering over charts and documents.
Mistral OCRmistral/mistral-ocr | Pixtralmistral/pixtral | |
|---|---|---|
| Type | OCR API | Vision model |
| Price | 25 credits / page | 2 credits / page |
| Per 1,000 pages | $25.00 | $2.00 |
| Typical speed | 0.9 s / page | 3.0 s / page |
| Best for | High-volume PDF to Markdown | Questions about document images |
| Inputs | PDF, images | images |
| Languages | 100+ languages | Multilingual |
| Max pages per request | 1,000 | 500 |
| Tables | ✓ Yes | No |
| Handwriting | No | No |
| Math | ✓ Yes | No |
| Bounding boxes | ✓ Yes | No |
| Confidence | No | No |
| Forms | No | No |
| Multilingual | ✓ Yes | ✓ Yes |
| On OCRRouter | Available | Available |
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
- It costs $2 per 1,000 pages, or $1 per 1,000 pages through the Batch API, which suits high-volume PDF to Markdown.
- OCR 3 rebuilds tables as HTML with rowspan and colspan, and it returns Markdown, HTML, JSON and extracted images.
- It is the current, maintained product, with a pinned version (mistral-ocr-2512), while Pixtral is marked deprecated.
Choose Pixtral if
- It ships as Apache-2.0 open weights, so you can self-host it and pay only for compute.
- It scores 90.7 ANLS on DocVQA (reported by Mistral), which fits question answering over charts and documents.
- Its 128K token context allows many images in one request, and it fits teams already integrated with the Mistral API.
Key differences
| Mistral OCR | Pixtral | |
|---|---|---|
| Model type | Dedicated document OCR model | General vision language model, 12.4B parameters (12B decoder + 400M vision encoder) |
| Licence | Proprietary | Apache-2.0 |
| How you run it | Hosted API only | Mistral API (pixtral-12b-2409) or self-hosted open weights |
| Price | $2 per 1,000 pages; $3 with structured annotations; $1 via Batch API | $0.15 per million input and output tokens at API launch; free to self-host |
| Output formats | Markdown, HTML, JSON, images | Plain text, Markdown, JSON |
| Benchmarks | OmniDocBench v1.5 overall 78.83 (first version); v1.6 overall 85.66 | DocVQA 90.7 ANLS, ChartQA 81.8, MMMU 52.5 (all reported by Mistral); no OmniDocBench score published |
| Known failures | Colored-background pages returned as images with no text (2025 release); hallucinated text on unreadable handwriting; changed checkbox states and misread characters on a dense financial table | Accuracy drops on handwriting, small fonts and non-Latin scripts; Q4 quantization degrades small text OCR; one HN user said it does not always get OCR near correct |
Sources
- Mistral OCR 3 announcement
- OmniDocBench v1.6 leaderboard
- Reducto test of Mistral OCR
- Mistral announcement
- HN launch thread
- Third-party deployment guide
What each one is
Mistral OCR
Fast document OCR that returns clean Markdown with tables and embedded images.
Mistral OCR detailsPixtral
Mistral's open-weight vision models, Pixtral 12B and Pixtral Large. Mistral's API no longer serves them, so they run self-hosted.
Pixtral detailsTry both on your own file
Run the file once with each model and compare the output.
Use one, fall back to the other
{"model": "mistral/mistral-ocr","fallbacks": ["mistral/pixtral"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Mistral OCR alternativesPixtral alternativesBest OCR models in 2026