OCRRouter

olmOCR vs dots.ocr

Pick olmOCR for large-scale PDF to Markdown, academic papers with equations, and pages that mix print and handwriting, and only if you have a GPU with at least 12 GB VRAM. Pick dots.ocr when you need layout boxes, categories and multilingual parsing from one smaller 3B model, with tables as HTML and formulas as LaTeX. On olmOCR-Bench the original dots.ocr scores 79.1 against 82.4 for olmOCR, though the newer dots.mocr reaches 83.9.

olmOCRallenai/olmocrdots.ocrrednote/dots-ocr
TypeOpen sourceOpen source
Price2 credits / page2 credits / page
Per 1,000 pages$2.00$2.00
Typical speed1.5 s / page1.5 s / page
Best forBulk PDF to text for datasetsMultilingual document parsing
InputsPDF, imagesPDF, images
LanguagesMainly English100+ languages
Max pages per request1,000500
Tables✓ Yes✓ Yes
Handwriting✓ YesNo
Math✓ Yes✓ Yes
Bounding boxesNo✓ Yes
ConfidenceNoNo
FormsNoNo
MultilingualNo✓ 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 olmOCR if

  • It scores 82.4 on olmOCR-Bench, above the original dots.ocr at 79.1.
  • Self-hosting costs under $200 per million pages per Ai2, and DeepInfra hosts olmOCR-2-7B at $0.09 input and $0.19 output per million tokens.
  • It handles equations, tables and handwriting, and strips headers and footers in natural reading order.

Choose dots.ocr if

  • One model returns layout boxes, categories and text, with tables as HTML and formulas as LaTeX.
  • It is a 3B model (1.7B LLM backbone) under an MIT-licensed repository, and its outputs include JSON, Markdown, HTML, LaTeX, bounding boxes and SVG.
  • The dots.mocr (v1.5) release reaches 83.9 on olmOCR-bench, and the dots.mocr-svg variant converts charts and diagrams to SVG.

Key differences

olmOCRdots.ocr
LicenceApache-2.0 (toolkit)MIT, with a separate dots.ocr license agreement file in the repository
How you run itSelf-host with vLLM on a GPU, or call third-party hosted endpointsSelf-host only, via vLLM or Hugging Face Transformers
Model size and hardware7B. NVIDIA GPU with at least 12 GB VRAM and 30 GB free disk3B (1.7B LLM backbone). Single CUDA GPU with vLLM is the documented path. No VRAM figure stated
PriceUnder $200 per million pages self-hosted. DeepInfra: $0.09 input / $0.19 output per million tokensFree software, you pay for compute
olmOCR-Bench82.4 (overall, v0.4.0)79.1 for dots.ocr, 83.9 for dots.mocr (1.5)
Output formatsMarkdown, JSONJSON, Markdown, HTML, LaTeX, bounding boxes, SVG
Known failureOut of memory on a 12 GB GPU (issue #424), can get stuck on the last few pages (issue #433), unreliable image bounding boxes (issue #437)Runs of ellipses or underscores can make output repeat endlessly. A table above 1000 grid cells raises an error and drops the whole page (Docling issue #5146)

Sources

What each one is

olmOCR

Open-source vision model from Ai2 (often searched as olm ocr) that turns PDFs into clean, reading-order text. Latest release: olmOCR 2.

olmOCR details

dots.ocr

Single vision model from rednote that does layout detection and content recognition together.

dots.ocr 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": "allenai/olmocr","fallbacks": ["rednote/dots-ocr"],"file_url": "https://example.com/document.pdf"}

Questions

More comparisons

olmOCR alternativesdots.ocr alternativesBest OCR models in 2026