OCRRouter

Jina OCR vs Chandra

Pick Chandra for multilingual scans, tables, forms and math, and for anything commercial, since its weights are free for companies under $2M funding or revenue. Pick Jina OCR for non-commercial PDF to Markdown and research, where it is light (about 570M active parameters) and scores 91.14 on OmniDocBench v1.6 by its authors' numbers. The main split is licence and scope: Jina OCR's CC BY-NC 4.0 blocks commercial use without a deal, while Chandra adds layout output and language coverage.

Jina OCRjina/jina-ocrChandradatalab/chandra
TypeOpen sourceOpen source
Price1 credit / page2 credits / page
Per 1,000 pages$1.00$2.00
Typical speed1.5 s / page2.0 s / page
Best forClean Markdown at low costComplex layouts and handwriting
InputsPDF, imagesPDF, images
LanguagesAbout 100 languages40+ languages
Max pages per request200500
Tables✓ Yes✓ Yes
HandwritingNo✓ Yes
Math✓ Yes✓ Yes
Bounding boxesNo✓ Yes
ConfidenceNoNo
FormsNo✓ Yes
Multilingual✓ Yes✓ Yes
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 Jina OCR if

  • You are doing research or non-commercial PDF to Markdown, where CC BY-NC 4.0 is not a problem.
  • You want a small active footprint: about 570M active parameters of a 3B MoE, at 2.57 pages per second on an A100.
  • You want Markdown, HTML or LaTeX output, with a GPU recommended and a CPU fallback in Transformers.

Choose Chandra if

  • You ship a commercial product and your company is under $2M funding or revenue, or you can buy a license above that.
  • Your documents are multilingual scans, or contain tables, forms and math. Chandra 2 averages 77.8% across 43 languages.
  • You need JSON output with layout chunks, bounding boxes and labels per block, or a hosted API at $1, $3 or $7.50 per 1,000 pages.

Key differences

Jina OCRChandra
LicenceCC-BY-NC-4.0. Commercial use needs a license from Jina/Elastic.Code Apache-2.0. Weights modified OpenRAIL-M, free for research, personal use and companies under $2M funding or revenue.
How you run itSelf-host or call Jina's hosted API.Self-host via vLLM or Hugging Face, or use the hosted Datalab API.
PriceFree self-hosted for non-commercial use, you pay for compute.Free self-hosted within licence terms. Datalab API is $1 (fast), $3 (balanced) or $7.50 (accurate) per 1,000 pages, with a $20/month free tier.
Output formatsMarkdown, HTML, LaTeX.Markdown, HTML, JSON, bounding boxes.
Benchmarks (authors' own)91.14 on OmniDocBench v1.6 and 83.4 on olmOCR-Bench.85.8 on olmOCR-bench (Chandra 2). The two olmOCR scores come from different authors' tables.
Hardware and speed3B MoE with about 570M active. 2.57 pages per second on an A100.One NVIDIA GPU via vLLM. About 1.44 pages/s on one H100 at 96 concurrent sequences. One reviewer measured 66 seconds per page on an RTX 4090 with the Hugging Face path.
Known failureHosted API returns HTTP 503 on cold start, so retry after 30 to 60 seconds. Reported scores use a stricter prompt than the recommended default.GitHub issue #51 reports token-loop hallucinations on 5-6 of about 60 Russian/Kazakh scans. Weak on Urdu (43.2%), Sanskrit (51.1%) and old scans (51.1 on olmOCR-bench).

Sources

What each one is

Jina OCR

Compact page-to-Markdown model from Jina AI, built on the DeepSeek-OCR encoder.

Jina OCR details

Chandra

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

Chandra 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": "jina/jina-ocr","fallbacks": ["datalab/chandra"],"file_url": "https://example.com/document.pdf"}

Questions

More comparisons

Jina OCR alternativesChandra alternativesBest OCR models in 2026