Gemini Flash vs dots.ocr
Pick Gemini Flash for cheap bulk PDF to Markdown or JSON extraction when you can send documents to a hosted API. Pick dots.ocr when documents must stay on your own hardware or you need layout boxes and categories from a free MIT-licensed model. The main split is hosted per-token billing versus a self-hosted 3B model where you pay for GPU compute.
Gemini Flashgoogle/gemini-flash | dots.ocrrednote/dots-ocr | |
|---|---|---|
| Type | Vision model | Open source |
| Price | 15 credits / page | 2 credits / page |
| Per 1,000 pages | $15.00 | $2.00 |
| Typical speed | 1.6 s / page | 1.5 s / page |
| Best for | Handwriting and mixed layouts | Multilingual document parsing |
| Inputs | PDF, images | PDF, images |
| Languages | 100+ languages | 100+ languages |
| Max pages per request | 1,000 | 500 |
| Tables | ✓ Yes | ✓ Yes |
| Handwriting | ✓ Yes | No |
| Math | ✓ Yes | ✓ Yes |
| Bounding boxes | No | ✓ 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 Gemini Flash if
- You want no infrastructure: send images or PDFs to the Gemini API or Vertex AI and prompt for Markdown or JSON.
- You need long documents in one request, since the context window goes up to 1M tokens.
- You want the stronger published score: Gemini 3 Flash reaches 92.62 on OmniDocBench v1.6, listed on the independent leaderboard.
Choose dots.ocr if
- You need on-prem or air-gapped deployment, because dots.ocr runs through vLLM or Hugging Face Transformers on your own machines.
- You want layout boxes, categories and text in one pass, with tables as HTML and formulas as LaTeX.
- You want no per-token fee and an MIT-licensed repository, with the 3B model priced only by your compute.
Key differences
| Gemini Flash | dots.ocr | |
|---|---|---|
| How you run it | Hosted API through Gemini API or Vertex AI | Self-hosted via vLLM or Hugging Face Transformers; single CUDA GPU with vLLM is the documented path |
| Licence | Proprietary | MIT, plus a separate dots.ocr license agreement file in the repository |
| Price | Gemini 3 Flash Preview: $0.50 per 1M input tokens, $3.00 per 1M output tokens; Batch/Flex $0.25 and $1.50 | Free to use; you pay for compute |
| Output formats | Markdown, JSON, HTML, plain text, shaped by your prompt with no fixed schema or confidence scores | JSON, Markdown, HTML, LaTeX, bounding boxes, SVG |
| Benchmark | 92.62 on OmniDocBench v1.6 (Gemini 3 Flash), independent | 79.1 on olmOCR-bench for dots.ocr and 83.9 for dots.mocr (v1.5), reported by the authors |
| Known failure modes | Cost per page varies with page density and output length; some newer Flash models rise from $0.75/$3.75 to $1.50/$7.50 per 1M tokens on January 1, 2027 | Runs of ellipses or underscores can make output repeat endlessly; complex tables and formulas stay difficult; a table above 1000 grid cells raised an error and dropped the page in Docling |
| Model size and hardware | Not published | 3B parameters (1.7B LLM backbone); no VRAM figure stated, and CPU-only production is not recommended |
Sources
- Gemini API pricing
- OmniDocBench leaderboard
- Gemini pricing changes 2026
- dots.ocr GitHub README
- dots.ocr blog / limitations
- Docling issue #5146
What each one is
Gemini Flash
General vision model with strong handwriting and layout understanding at a low price.
Gemini Flash detailsdots.ocr
Single vision model from rednote that does layout detection and content recognition together.
dots.ocr 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": "google/gemini-flash","fallbacks": ["rednote/dots-ocr"],"file_url": "https://example.com/document.pdf"}Questions
More comparisons
Gemini Flash alternativesdots.ocr alternativesBest OCR models in 2026