Frontier AI forDocument Extraction

NuMind builds state-of-the-art VLMs specialized in extracting information from documents. Our models power our SaaS platform and are privately deployed across insurance, banking, healthcare, and other document-intensive industries.

Deployed inside the most demanding companies

30%97%

Extraction accuracy after switching to NuExtract

AXA’s previous automated workflow achieved 30% field-level accuracy, requiring extensive human review. With NuExtract, accuracy increased to 97% (a 23-fold reduction in errors), dramatically lowering manual validation and extraction costs.

30%97%

Extraction accuracy after switching to NuExtract

AXA’s previous automated workflow achieved 30% field-level accuracy, requiring extensive human review. With NuExtract, accuracy increased to 97% (a 23-fold reduction in errors), dramatically lowering manual validation and extraction costs.

In-house models, validated at open-source scale

5M+Downloads on Hugging Face

Explore our models

Our models lead the benchmarks

NuExtract outperforms generalist models for both structured extraction and OCR.

Structured extraction benchmark: NuExtract 3 outperforms similarly sized models.

NuExtract3 scores 65.2, compared with 53.8 for the next-best similarly sized model.

Research that advances
the field

Featured publication

EMNLP 2024

100+ citations

NuNER

Entity Recognition Encoder Pre-training via LLM-Annotated Data

MACHINE LEARNING

BLOG

NuExtract3

The Reasoning Open-Source OCR & Structured Extraction LLM.