Pixtral vs Chunkr
Pick Chunkr if you need to parse PDFs, slides and Word files into HTML, Markdown or JSON with bounding boxes for a RAG pipeline. Pick Pixtral only if you already run on the Mistral API and want question answering over charts and documents. Pixtral 12B is deprecated and no longer maintained, so it is a poor base for a new OCR project.
Pixtralmistral/pixtral | Chunkrchunkr/chunkr | |
|---|---|---|
| Type | Vision model | OCR API |
| Price | 2 credits / page | 14 credits / page |
| Per 1,000 pages | $2.00 | $14.00 |
| Typical speed | 3.0 s / page | 3.0 s / page |
| Best for | Questions about document images | Layout-aware chunks for RAG |
| Inputs | images | PDF, images, DOCX |
| Languages | Multilingual | Multilingual |
| Max pages per request | 500 | 1,000 |
| Tables | No | ✓ Yes |
| Handwriting | No | No |
| Math | No | ✓ Yes |
| Bounding boxes | No | ✓ Yes |
| 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 Pixtral if
- You already have an integration on the Mistral API and want to keep it.
- You need question answering over charts and documents, where Mistral reports 90.7 ANLS on DocVQA and 81.8 on ChartQA.
- You want Apache-2.0 open weights with no licence friction for closed products.
Choose Chunkr if
- You need layout analysis, OCR with bounding boxes and semantic chunking for RAG over PDFs and slides.
- You want structured output in HTML, Markdown or JSON from PDF, PPT, Word and image inputs.
- You want to self-host with Docker on GPU, CPU-only or Mac ARM, using any OpenAI-compatible LLM endpoint.
Key differences
| Pixtral | Chunkr | |
|---|---|---|
| What it is | General vision language model from Mistral AI: 12B Mistral Nemo decoder plus a 400M vision encoder | Document parsing service from Lumina AI: layout analysis, OCR with bounding boxes, semantic chunking |
| Licence | Apache-2.0 | AGPL-3.0 or commercial licence; closed-source products need the commercial licence |
| How you run it | Mistral API (pixtral-12b-2409) or open weights you host yourself | Cloud API or self-hosted open-source version; self-host still needs an external LLM endpoint |
| Price | $0.15 per million input and output tokens at launch; free to self-host, you pay for compute | Credit-based hosted plans, prices not published; open-source version is free, you pay for compute and an LLM endpoint |
| Output formats | Plain text, Markdown, JSON | HTML, Markdown, JSON, bounding boxes |
| Benchmarks | DocVQA 90.7 ANLS, ChartQA 81.8, MMMU 52.5 (reported by Mistral) | Not published |
| Known failure | Reported to miss on handwriting, small fonts and non-Latin scripts; Q4 quantization degrades small text; deprecated | Open-source version uses community models that are less accurate than the cloud's; no Excel support in the open-source version |
Sources
- Mistral announcement
- Mistral Sept 2024 pricing
- Third-party deployment guide
- Chunkr GitHub README
- Chunkr site
What each one is
Pixtral
Mistral's open-weight vision models, Pixtral 12B and Pixtral Large. Mistral's API no longer serves them, so they run self-hosted.
Pixtral detailsChunkr
Document parsing API that segments and chunks pages for RAG. Self-hostable version available.
Chunkr 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/pixtral","fallbacks": ["chunkr/chunkr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Pixtral alternativesChunkr alternativesBest OCR models in 2026