Florence-2 vs Chunkr
Pick Florence-2 for short English text in scenes or labels where you need line boxes, or when one small model must also do captioning and detection. Pick Chunkr for full PDFs, slides and Word files that need Markdown, HTML or JSON for RAG chunking. The main split is output: Florence-2 returns only text and boxes, while Chunkr parses document layout.
Florence-2microsoft/florence-2 | Chunkrchunkr/chunkr | |
|---|---|---|
| Type | Open source | OCR API |
| Price | 1 credit / page | 14 credits / page |
| Per 1,000 pages | $1.00 | $14.00 |
| Typical speed | 0.8 s / page | 3.0 s / page |
| Best for | Lightweight OCR with boxes | Layout-aware chunks for RAG |
| Inputs | images | PDF, images, DOCX |
| Languages | English | Multilingual |
| Max pages per request | 500 | 1,000 |
| Tables | No | ✓ Yes |
| Handwriting | No | No |
| Math | No | ✓ Yes |
| Bounding boxes | ✓ Yes | ✓ Yes |
| Confidence | No | No |
| Forms | No | No |
| Multilingual | No | ✓ 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 Florence-2 if
- It is a 0.23B (base) or 0.77B (large) model under the MIT licence, so it is cheap to self-host and has no copyleft terms.
- OCR_WITH_REGION returns quadrilateral boxes per text line.
- The same model also does captioning, detection and grounding, so one model can cover several vision tasks in a pipeline.
Choose Chunkr if
- It outputs HTML, Markdown, JSON and OCR with bounding boxes, which suits RAG chunking of PDFs and slides.
- It accepts PDF, PPT, Word and images, so you do not have to convert files before parsing.
- You can use the hosted API or self-host with Docker (GPU, CPU-only or Mac ARM), with any OpenAI-compatible LLM endpoint.
Key differences
| Florence-2 | Chunkr | |
|---|---|---|
| Licence | MIT | AGPL-3.0 or commercial licence; closed-source use needs the commercial licence |
| How you run it | Self-host via Transformers only | Cloud API or self-host via Docker; self-host still needs an external LLM endpoint |
| Price | Free, self-hosted (you pay for compute) | Hosted credit-based plans, official prices not verified; open-source version free (you pay for compute and an LLM endpoint) |
| Output formats | Plain text and bounding boxes; no Markdown, tables or LaTeX | HTML, Markdown, JSON and bounding boxes |
| Input | Images (scene text, short labels) | PDF, PPT, Word and images; Excel not supported in the open-source version |
| Languages and handwriting | Chinese output mostly wrong because the vocabulary lacks CJK characters; spelling errors and missing spaces on cursive handwriting and French text in an HF blog test | Not published |
| Known weakness | No document structure output; the continued-pretrained checkpoint used only 0.1B samples and might not be well trained | Open-source version uses community models, less accurate than the cloud's in-house models |
Sources
What each one is
Florence-2
Small Microsoft vision model (0.23B and 0.77B) with OCR and OCR-with-regions tasks that return text with boxes.
Florence-2 detailsChunkr
Document parsing API that segments and chunks pages for RAG. Self-hostable version available.
Chunkr detailsTry both on your own file
Run the file once with each model and compare the output.
Use one, fall back to the other
{"model": "microsoft/florence-2","fallbacks": ["chunkr/chunkr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Florence-2 alternativesChunkr alternativesBest OCR models in 2026