Florence-2 vs Pixtral
Pick Pixtral for question answering over charts and documents if you already run on the Mistral API, but note that Mistral marks it deprecated. Pick Florence-2 for small, self-hosted English OCR where you need bounding boxes per text line. Neither suits handwriting or non-Latin scripts.
Florence-2microsoft/florence-2 | Pixtralmistral/pixtral | |
|---|---|---|
| Type | Open source | Vision model |
| Price | 1 credit / page | 2 credits / page |
| Per 1,000 pages | $1.00 | $2.00 |
| Typical speed | 0.8 s / page | 3.0 s / page |
| Best for | Lightweight OCR with boxes | Questions about document images |
| Inputs | images | images |
| Languages | English | Multilingual |
| Max pages per request | 500 | 500 |
| Tables | No | No |
| Handwriting | No | No |
| Math | No | No |
| Bounding boxes | ✓ Yes | No |
| 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 tiny at 0.23B (base) or 0.77B (large), so it fits edge and low-cost hardware.
- OCR_WITH_REGION returns quadrilateral boxes per text line, which suits scene text and short labels.
- One model also covers captioning, detection and grounding, so one pipeline can do all of them.
Choose Pixtral if
- It reports 90.7 ANLS on DocVQA and 81.8 on ChartQA, so it fits question answering over documents and charts.
- It outputs plain text, Markdown and JSON, while Florence-2 gives only text and boxes.
- It is available through the Mistral API at a launch price of $0.15 per million tokens, or as Apache-2.0 open weights.
Key differences
| Florence-2 | Pixtral | |
|---|---|---|
| Licence | MIT | Apache-2.0 |
| How you run it | Self-host only, via Transformers | Mistral API (pixtral-12b-2409) or self-host open weights |
| Size | 0.23B (base), 0.77B (large) | 12.4B (12B decoder + 400M vision encoder) |
| Price | Free, you pay for compute | $0.15 per million input and output tokens at API launch; free to self-host |
| Output formats | Plain text and bounding boxes | Plain text, Markdown and JSON |
| Document benchmark | Not published (TextCaps CIDEr 72.8 zero-shot, large) | DocVQA 90.7 ANLS, ChartQA 81.8 (CoT) |
| Known failures | Spelling errors and missing spaces on cursive handwriting and French text; Chinese output mostly wrong | Does not always get OCR near correct; accuracy drops on handwriting, small fonts and non-Latin scripts; Q4 quantization degrades small text |
Sources
- Florence-2-large model card
- HF discussion Chinese very bad
- Mistral announcement
- Mistral Sept 2024 pricing
- HN launch thread
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 detailsPixtral
Mistral's open-weight vision models, Pixtral 12B and Pixtral Large. Mistral's API no longer serves them, so they run self-hosted.
Pixtral 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": ["mistral/pixtral"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Florence-2 alternativesPixtral alternativesBest OCR models in 2026