OCRRouter

LlamaParse vs Hunyuan OCR

Pick LlamaParse if you want a hosted API for PDFs and office files feeding a RAG pipeline, and you can tolerate async latency. Pick Hunyuan OCR if you want to run a ~1B model on your own hardware, or need photo text spotting, image translation or low-resource scripts. The main split is hosted per-page billing versus self-hosting with a regional licence limit.

LlamaParsellamaindex/llamaparseHunyuan OCRtencent/hunyuan-ocr
TypeOCR APIOpen source
Price5 credits / page2 credits / page
Per 1,000 pages$5.00$2.00
Typical speed3.0 s / page1.5 s / page
Best forRAG ingestionChinese documents and photos
InputsPDF, images, DOCXPDF, images
Languages80+ languages100+ languages
Max pages per request1,000500
Tables✓ Yes✓ Yes
HandwritingNoNo
Math✓ Yes✓ Yes
Bounding boxes✓ Yes✓ Yes
ConfidenceNoNo
FormsNoNo
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 LlamaParse if

  • You want a hosted API with no GPU or CUDA setup, and 10,000 free credits per month to start.
  • You parse mixed PDF and office files into Markdown or JSON for RAG inside LlamaIndex pipelines.
  • You want to pay by difficulty: from $1.25 per 1,000 pages in Fast mode up to agentic modes for hard layouts.

Choose Hunyuan OCR if

  • You want to self-host a ~1B model, including on a laptop through the llama.cpp GGUF path.
  • You need photo and scene text spotting, text-image translation, or low-resource and ancient-script OCR in one model.
  • You want bounding boxes plus Markdown, HTML, LaTeX or JSON output, with images up to 4K and a 128K context window in v1.5.

Key differences

LlamaParseHunyuan OCR
How you run itHosted API from LlamaIndex (LlamaCloud)Self-hosted via vLLM, Transformers or llama.cpp
LicenceProprietaryTencent Hunyuan Community License. Excludes the EU, UK and South Korea. Separate licence needed above 100M monthly active users
Price$1.25 per 1,000 credits. Fast is 1 credit per page, Cost Effective 3, Agentic 10, Agentic Plus 45. 10K free credits per monthFree software. You pay for your own compute
Output formatsMarkdown, JSON, plain text, imagesMarkdown, HTML, LaTeX, JSON, plain text, bounding boxes
Setup requirementsAPI access onlyNVIDIA GPU with CUDA 13 and vLLM 0.25.1+ for the unified environment, or llama.cpp GGUF for CPU, consumer GPU or laptop
BenchmarkNot publishedParseBench-2 overall 67.11 and ParseBench mean 63.86 (v1.5, independent)
Known failureInvented transactions on bank statements in Accurate mode (GitHub #450). Hallucination on low-quality scans with no confidence score (#483). A job stuck in progress for days (#623)Transformers and vLLM outputs differed in accuracy in the archived 1.0 note. Text score of 62.88 in the public CI replay

Sources

What each one is

LlamaParse

Document parser built for RAG pipelines, with several quality tiers.

LlamaParse details

Hunyuan OCR

Small end-to-end OCR model from Tencent covering parsing, extraction and photo translation.

Hunyuan 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": "llamaindex/llamaparse","fallbacks": ["tencent/hunyuan-ocr"],"file_url": "https://example.com/document.pdf"}

Questions

More comparisons

LlamaParse alternativesHunyuan OCR alternativesBest OCR models in 2026