LlamaParse vs Marker
Pick LlamaParse if you want a hosted API for mixed PDF and office files inside a LlamaIndex RAG pipeline, with a free tier and per-page tiers. Pick Marker if you want to run PDF to Markdown on your own hardware, or if you handle books and papers with equations. The main split is hosted and proprietary versus self-hosted and open code, and Marker's model weights carry a licence limit for companies above $5M.
LlamaParsellamaindex/llamaparse | Markerdatalab/marker | |
|---|---|---|
| Type | OCR API | OCR API |
| Price | 5 credits / page | 14 credits / page |
| Per 1,000 pages | $5.00 | $14.00 |
| Typical speed | 3.0 s / page | 2.5 s / page |
| Best for | RAG ingestion | Books and research papers |
| Inputs | PDF, images, DOCX | PDF, images, DOCX |
| Languages | 80+ languages | 90+ languages |
| Max pages per request | 1,000 | 1,000 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | No | No |
| Math | ✓ Yes | ✓ Yes |
| Bounding boxes | ✓ Yes | ✓ Yes |
| Confidence | No | No |
| Forms | No | 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 LlamaParse if
- It parses PDFs and office files, so mixed document sets go through one service.
- You pay per page by tier, from 1 credit per page in Fast mode up to 45 in Agentic Plus, with 10,000 free credits per month.
- Re-parsing the same file within 48 hours is free because results are cached.
Choose Marker if
- You can run it locally on GPU, CPU or Apple Silicon, so documents stay on your own machines.
- It outputs Markdown, HTML, JSON block trees, RAG chunks and LaTeX, which suits books and papers with equations.
- Fast mode runs on CPU at 23.6 pages per second with about 18 MB of ONNX weights, and self-hosting costs only your compute.
Key differences
| LlamaParse | Marker | |
|---|---|---|
| How you run it | Hosted API only | Self-hosted on GPU, CPU or MPS; Datalab also sells a hosted API |
| Licence | Proprietary | Code Apache-2.0; weights under a modified AI Pubs Open Rail-M license, free for research, personal use and startups under $5M |
| Price | $1.25 per 1,000 credits. Fast 1 credit/page, Cost Effective 3, Agentic 10, Agentic Plus 45. 10K free credits per month | Free self-hosted within the weight license, you pay for compute. Hosted API has about 5,000 free pages per month |
| Output formats | Markdown, JSON, plain text, images | Markdown, HTML, JSON, chunks, LaTeX, images |
| Benchmarks | Not published | 83.7 on olmOCR-bench and 85.9 on OmniParseBench in vllm mode, both reported by Datalab; CPU fast mode scores 66.6 on olmOCR-bench |
| Known failure | GitHub issues report invented transactions on bank statements and hallucination on low-quality scans with no confidence score | Weight license blocks commercial use for companies above $5M without a paid license |
| Latency | GitHub issue reports a job stuck in progress for days and premium mode taking 4 to 5 minutes | Fast mode on CPU runs at 23.6 pages per second |
Sources
- LlamaIndex pricing
- GitHub issue #450
- GitHub issue #483
- GitHub issue #623
- Marker GitHub README
- Datalab pricing
What each one is
Marker
Hosted version of the open-source Marker and Surya models. Strong on books and papers.
Marker 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": "llamaindex/llamaparse","fallbacks": ["datalab/marker"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
LlamaParse vs Amazon TextractLlamaParse vs Reducto ParseMathpix Convert vs MarkerClaude vs MarkerLlamaParse vs DoclingMarker vs docTRGPT vs LlamaParseLlamaParse vs MonkeyOCR
LlamaParse alternativesMarker alternativesBest OCR models in 2026