Nanonets-OCR2-3B – OCR model that transforms documents into structured markdown
2 days ago
- #AI
- #markdown
- #OCR
- Nanonets-OCR2 is a state-of-the-art OCR model family that converts documents into structured markdown with semantic tagging.
- Features include handling complex documents, recognizing equations, images, signatures, and watermarks, and tagging them for LLM processing.
- Available models include Nanonets-OCR2-Plus, Nanonets-OCR2-3B, and Nanonets-OCR2-1.5B-exp, with performance comparisons provided.
- The model can be used via Python code snippets or API calls, supporting various document types including financial documents.
- Performance benchmarks show Nanonets-OCR2 models competing with other leading models like Gemini and GPT-5.