# LightOn Releases OCR-2, a 1B Document AI Model

> The new vision model from the Paris-based AI lab uses Mistral architecture to extract text and structure from complex documents like PDFs and forms.

Published by The Open Weights on Jan 16, 2026. Canonical: https://theopenweights.com/news/lightonocr-2-1b-rre8

## Key facts

- Company: LightOn
- Model: LightOnOCR-2 1B
- Version: 2
- Category: Vision-Language
- Modalities: Vision-Language
- License: Other (Open weights, commercial use allowed)
- Parameters: 1B
- Context window: 16K tokens
- Architecture: LightOnOCRForConditionalGeneration
- Precision: BF16
- Weights size: 2 GB
- Significance: notable
- Published: 2026-01-16
- Last verified: 2026-07-30
- Hugging Face: https://huggingface.co/lightonai/LightOnOCR-2-1B
- Canonical URL: https://theopenweights.com/news/lightonocr-2-1b-rre8

Parisian AI company LightOn has released LightOnOCR-2, a new 1-billion-parameter vision language model specialized in document understanding. The model is designed to perform optical character recognition (OCR) on complex documents, extracting not just text but also structural information.

Unlike simple OCR tools, LightOnOCR-2 is built to parse challenging layouts like tables, forms, and multi-column PDFs. This makes it suitable for enterprise automation tasks such as processing invoices or digitizing records, a domain often dominated by proprietary, API-gated services.

## A Mistral-based Architecture

The model features a Transformer-based encoder-decoder architecture. In an interesting design choice, its decoder was initialized using a subset of weights from Mistral-7B-v0.1, allowing it to leverage the powerful language capabilities of the popular open model while maintaining a much smaller, more efficient footprint.

The complete model weights and code are available on the [Hugging Face Hub](https://huggingface.co/lightonai/LightOnOCR-2-1B) for developers to download and use. It's released under a custom LightOnAI-OpenRAIL-M license, which permits commercial use but includes some use-case restrictions common to Responsible AI licenses.

## Get the model

- [Hugging Face](https://huggingface.co/lightonai/LightOnOCR-2-1B)

## Sources

- [lightonai/LightOnOCR-2-1B](https://huggingface.co/lightonai/LightOnOCR-2-1B) — Hugging Face, Jan 16, 2026

---
Source: The Open Weights (https://theopenweights.com/). Aggregated and written by Claude, curated by humans. Cite as: "LightOn Releases OCR-2, a 1B Document AI Model", The Open Weights, Jan 16, 2026, https://theopenweights.com/news/lightonocr-2-1b-rre8