Mistral OCR vs olmOCR
Pick Mistral OCR if you want a hosted API with no GPU to run and a list price of $2 per 1,000 pages, or $1 with the Batch API. Pick olmOCR if you can run an NVIDIA GPU and want open weights, Apache-2.0 tooling and a self-hosted cost under $200 per million pages. On olmOCR's own bench it scores 82.4 against 72.0 for Mistral OCR, but that table is published by olmOCR's authors.
Mistral OCRmistral/mistral-ocr | olmOCRallenai/olmocr | |
|---|---|---|
| Type | OCR API | Open source |
| Price | 25 credits / page | 2 credits / page |
| Per 1,000 pages | $25.00 | $2.00 |
| Typical speed | 0.9 s / page | 1.5 s / page |
| Best for | High-volume PDF to Markdown | Bulk PDF to text for datasets |
| Inputs | PDF, images | PDF, images |
| Languages | 100+ languages | Mainly English |
| Max pages per request | 1,000 | 1,000 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | No | ✓ Yes |
| Math | ✓ Yes | ✓ Yes |
| Bounding boxes | ✓ Yes | No |
| Confidence | No | No |
| Forms | No | No |
| Multilingual | ✓ Yes | No |
| 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
- You want a hosted API with no GPU to provision, at $2 per 1,000 pages or $1 per 1,000 with the Batch API.
- You need tables rebuilt as HTML with rowspan and colspan, plus extracted images, in the output.
- You process high-volume scanned business documents or PDF to Markdown for RAG in batch.
Choose olmOCR if
- You can run an NVIDIA GPU with at least 12 GB VRAM and want self-hosted cost under $200 per million pages, per Ai2.
- You want open weights, training data and code, with the toolkit under Apache-2.0.
- Your pages are academic papers with equations, or mixed printed and handwritten pages, and you want headers and footers stripped in reading order.
Key differences
| Mistral OCR | olmOCR | |
|---|---|---|
| Deployment | Hosted API only (mistral-ocr-2512, OCR 3) | Self-host with vLLM on a GPU, or use third-party hosted endpoints |
| Licence | Proprietary | Apache-2.0 (toolkit); weights, training data and code released |
| Price | $2 per 1,000 pages; $3 with structured annotations; $1 with Batch API | Under $200 per million pages self-hosted (Ai2); DeepInfra $0.09 input / $0.19 output per million tokens |
| Output formats | Markdown, HTML, JSON, images | Markdown, JSON |
| Benchmark | OmniDocBench v1.6 overall: 85.66 (independent); v1.5 for the first version: 78.83 | olmOCR-Bench v0.4.0 overall: 82.4 (reported by its authors), versus 72.0 for Mistral OCR in the same table |
| Hardware | None to manage | NVIDIA GPU with at least 12 GB VRAM and 30 GB free disk; no CPU-only use |
| Known failure | Hallucinated text on unreadable handwriting; colored-background pages returned as images with no OCR text (2025 release); changed checkbox states on a dense financial table | Out of memory on a 12 GB GPU; gets stuck on the last few pages; unreliable image bounding box coordinates |
Sources
- Mistral OCR 3 announcement
- VentureBeat on OCR 3 pricing
- OmniDocBench v1.6 leaderboard
- olmOCR GitHub README
- Issue #424 OOM
- Reducto test of Mistral OCR
What each one is
Mistral OCR
Fast document OCR that returns clean Markdown with tables and embedded images.
Mistral OCR detailsolmOCR
Open-source vision model from Ai2 (often searched as olm ocr) that turns PDFs into clean, reading-order text. Latest release: olmOCR 2.
olmOCR 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": ["allenai/olmocr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Mistral OCR alternativesolmOCR alternativesBest OCR models in 2026