OCRRouter

Chandra vs Nougat

Pick Chandra for mixed, multilingual or business documents, and for any commercial product. Pick Nougat only for English arXiv-style papers where you want LaTeX math, and only for non-commercial work. The main split is licence and scope: Nougat's weights are CC-BY-NC and it is built for academic papers, while Chandra covers tables, forms and 43+ languages.

Chandradatalab/chandraNougatmeta/nougat
TypeOpen sourceOpen source
Price2 credits / page2 credits / page
Per 1,000 pages$2.00$2.00
Typical speed2.0 s / page3.0 s / page
Best forComplex layouts and handwritingarXiv-style papers
InputsPDF, imagesPDF, images
Languages40+ languagesMainly English
Max pages per request500200
Tables✓ Yes✓ Yes
Handwriting✓ YesNo
Math✓ Yes✓ Yes
Bounding boxes✓ YesNo
ConfidenceNoNo
Forms✓ YesNo
Multilingual✓ YesNo
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 Chandra if

  • You need a hosted option: the Datalab API charges $1, $3 or $7.50 per 1,000 pages, with a $20/month free tier.
  • Your documents are scanned, multilingual or contain forms and tables, since Chandra 2 averages 77.8% across 43 languages.
  • You need layout data, because it returns Markdown, HTML or JSON plus bounding boxes and labels per block.

Choose Nougat if

  • Your input is English scientific papers, which is what it was trained on (arXiv and PubMed Central).
  • You want equations and tables as LaTeX in Mathpix-compatible .mmd Markdown.
  • You want a small self-hosted Python package with small and base checkpoints, a simple CLI and an API, for non-commercial use.

Key differences

ChandraNougat
LicenceCode Apache-2.0; weights modified OpenRAIL-M, free under $2M funding/revenue, commercial licence above thatCode MIT; weights CC-BY-NC, no commercial use
How you run itSelf-host with vLLM or Hugging Face, or use the hosted Datalab APISelf-hosted Python package only
PriceSelf-hosted: free within licence terms. API: $1, $3 or $7.50 per 1,000 pagesFree for non-commercial use, you pay for compute
Output formatsMarkdown, HTML, JSON, bounding boxesMarkdown (.mmd) and LaTeX
Languages77.8% average over 43 languages; weak on Urdu (43.2%) and Sanskrit (51.1%)Works best on English; Chinese, Russian and Japanese do not work
Published benchmarkChandra 2: 85.8 on olmOCR-bench (authors' table)Not published
Known failureToken-loop hallucinations on 5-6 of about 60 Russian/Kazakh scans in one issue; weak on old scans (51.1 on olmOCR-bench)Repetition loops on 1.5% of test pages, more out of domain; looping pages can be dropped as MISSING_PAGE

Sources

What each one is

Chandra

Datalab's OCR model that keeps layout and handles forms, tables and handwriting.

Chandra details

Nougat

Meta's model that converts academic PDFs to Markdown with LaTeX math.

Nougat 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": "datalab/chandra","fallbacks": ["meta/nougat"],"file_url": "https://example.com/document.pdf"}

Questions

More comparisons

Chandra alternativesNougat alternativesBest OCR models in 2026