TAAFT
Free mode
100% free
Freemium
Free Trial
Deals

NuExtract3

By NuMind
NuExtract3 is a 4B multimodal document-understanding model from NuMind, released under Apache 2.0. It accepts text, images, or text-plus-image inputs and is built for structured information extraction, OCR-style document parsing, document-to-Markdown conversion, RAG preprocessing, multilingual documents, and template generation. Its main workflow is document input plus JSON template plus optional instructions, producing structured JSON output, while its Markdown mode converts document images into clean Markdown with tables and layout preserved where possible. It also supports reasoning and non-reasoning inference modes, making it useful for both fast deterministic extraction and harder document-understanding tasks.
New Multimodal Gen 3
Released: May 12, 2026

Overview

NuExtract3 is NuMind’s open-source 4B vision-language reasoning model for document understanding, structured extraction, OCR, and document-to-Markdown conversion.

About NuMind

NuMind is an Artificial Intelligence company that enables software engineers, data scientists, and non-experts to effortlessly create advanced machine learning models utilizing LLMs for automated text processing.

Industry: Software Development
Company Size: 9
Location: Wilmington, Delaware, US
Website: numind.ai
View Company Profile

Tools using NuExtract3

No tools found for this model yet.

Last updated: May 26, 2026
0 AIs selected
Clear selection
#
Name
Task