Chandra vs Pixtral
Pick Chandra for new OCR work: scanned pages, tables, forms and multilingual documents that need Markdown, HTML or JSON with layout. Pixtral 12B is a general vision model that Mistral now marks as deprecated, so keep it only for existing Mistral API integrations or question answering over charts and documents.
Chandradatalab/chandra | Pixtralmistral/pixtral | |
|---|---|---|
| Type | Open source | Vision model |
| Price | 2 credits / page | 2 credits / page |
| Per 1,000 pages | $2.00 | $2.00 |
| Typical speed | 2.0 s / page | 3.0 s / page |
| Best for | Complex layouts and handwriting | Questions about document images |
| Inputs | PDF, images | images |
| Languages | 40+ languages | Multilingual |
| Max pages per request | 500 | 500 |
| Tables | ✓ Yes | No |
| Handwriting | ✓ Yes | No |
| Math | ✓ Yes | No |
| Bounding boxes | ✓ Yes | No |
| Confidence | No | No |
| Forms | ✓ Yes | 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 Chandra if
- It scores 85.8 on olmOCR-bench and averages 77.8% across 43 languages, so it is built for transcribing pages, not just answering questions about them.
- It returns Markdown, HTML or JSON plus layout chunks with bounding boxes and labels per block, which suits PDF to Markdown pipelines for RAG.
- You can self-host on a single GPU through vLLM, or use the hosted Datalab API at $1, $3 or $7.50 per 1,000 pages.
Choose Pixtral if
- It has Apache-2.0 weights with no revenue cap, while Chandra's weights need a commercial license for companies above $2M funding or revenue.
- It scores 90.7 ANLS on DocVQA and 81.8 on ChartQA, which fits question answering over documents and charts.
- If you already call the Mistral API, it costs $0.15 per million input and output tokens and has a 128K token context for many images per request.
Key differences
| Chandra | Pixtral | |
|---|---|---|
| Licence | Code Apache-2.0; weights modified OpenRAIL-M, free for research, personal use and companies under $2M funding/revenue | Apache-2.0 |
| Maintenance status | Current; Chandra 2 is the model benchmarked | Marked deprecated by Mistral and no longer maintained |
| Price | Self-hosted is free within licence terms. Datalab API is $1, $3 or $7.50 per 1,000 pages, with a $20/month free tier | Mistral API launch price $0.15 per million input and output tokens; free to self-host |
| Output formats | Markdown, HTML, JSON and bounding boxes | Plain text, Markdown and JSON |
| Benchmarks | olmOCR-bench 85.8; 77.8% average over 43 languages | DocVQA 90.7 ANLS; ChartQA 81.8; MMMU 52.5 |
| Handwriting and non-Latin text | Weak on low-resource scripts: 43.2% Urdu, 51.1% Sanskrit. Old Scans only 51.1 on olmOCR-bench | Reported accuracy drop on handwriting, small fonts and non-Latin scripts |
| Known failure | Token-loop hallucinations on 5-6 of about 60 Russian/Kazakh scans in one GitHub issue; Datalab says raw output hallucinates on about 6 in 10 of its hardest pages | A launch thread user says it does not always get OCR anywhere near correct; Q4 quantization reported to degrade small text OCR |
Sources
- Chandra GitHub README
- Datalab pricing
- GitHub issue #51 token loops
- Mistral announcement
- HN launch thread
What each one is
Chandra
Datalab's OCR model that keeps layout and handles forms, tables and handwriting.
Chandra 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.
PDF, PNG, JPG, WebP, HEIC. Up to 20 MB each, 10 files.No file?
Use one, fall back to the other
{"model": "datalab/chandra","fallbacks": ["mistral/pixtral"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Chandra vs InternVL3.5Chandra vs RolmOCRSurya vs PixtralolmOCR vs PixtralMistral OCR vs PixtralMathpix Convert vs PixtralTrOCR vs ChandraChandra vs Tesseract.js
Chandra alternativesPixtral alternativesBest OCR models in 2026