Skip to content
AI Model Radar

NVIDIA Nemotron Parse 2.0

nvidia/NVIDIA-Nemotron-Parse-2.0
Open Source · chat · open-weights
Open Alert me on changes
Context
—
Max output
—
Weights
Open
API $/1M
—
Modalities
text · image
Released
30 Jun 2026
License: other · nvidia/NVIDIA-Nemotron-Parse-2.0
Download image Share on X Share on LinkedIn
AI summary
● machine-written

NVIDIA Nemotron Parse 2.0 released for document intelligence

NVIDIA Nemotron Parse 2.0 is an open-weights vision-language model designed for document understanding and semantic parsing. It extracts text and table elements with spatial grounding, overcoming traditional OCR limitations through support for multi-column layouts, LaTeX extraction, markdown formatting, and reading-order reconstruction. The model is optimized for document intelligence pipelines including RAG, LLM training data curation, and agentic document workflows.

What's new
  • Open-weights model available on Hugging Face
  • Handles multi-column layouts and LaTeX table extraction
  • Supports markdown formatting and reading-order reconstruction
  • Deployed as NVIDIA NIM API microservice
  • Compatible with vLLM, SGLang, and Transformers libraries
Best for
Document intelligence and OCR tasksRAG pipeline data extractionLLM training data curationAgentic document workflows
Sources

Source: https://huggingface.co/nvidia/NVIDIA-Nemotron-Parse-2.0