LightOnOCR-2-1B

Summary: LightOnOCR-2-1B is a compact 1B-parameter end-to-end multilingual vision-language model developed by LightOn. It converts document images, such as scans and photos, into clean, naturally ordered text while handling complex layouts, including tables, multi-column documents, and scientific notation. It is available through the IONOS CLOUD OpenAI-compatible API for production-ready OCR.

The model accepts image input only. PDFs must be converted to images before being sent to the API.

Intelligence	Speed	Sovereignty	Input	Output
Moderate	High	High	Image	Text

Central parameters

Description: 1B-parameter end-to-end vision-language model for document OCR. Converts scans and images into clean, naturally ordered Markdown text. Handles complex layouts including tables, multi-column documents, scientific notation with LaTeX, and scanned material. The model always outputs Markdown-formatted text; text prompts are accepted by the API but do not influence the output format. The model accepts image input only. PDFs must be converted to images before being sent to the API.

Model identifier: lightonai/LightOnOCR-2-1B

IONOS CLOUD AI Model Hub Lifecycle and Alternatives

IONOS start date	End of Life	Alternative	Successor
February 23, 2026	N/A	—	—

Origin

Provider	Country	License	Flavor	Release
LightOn	France	Apache 2.0	Base	2026

Technology

Context window	Parameters	Architecture	Multilingual	Further details
16k	1B	Vision-Language Model	Yes	Hugging Face

Modalities

Text	Image	Audio
Output	Input	Not supported

Endpoints

Chat Completions	Embeddings	Image generation
v1/chat/completions	Not supported	Not supported

OCR Example

{
  "model": "lightonai/LightOnOCR-2-1B",
  "messages": [
    {
      "role": "user",
      "content": [
        {
          "type": "image_url",
          "image_url": {
            "url": "https://example.com/document.png"
          }
        }
      ]
    }
  ],
  "max_tokens": 4096,
  "temperature": 0.2,
  "top_p": 0.9
}

Note: LightOnOCR-2-1B accepts image input only. PDF input is not supported. If your source document is a PDF, convert each page to an image (for example, PNG or JPEG) before sending it to the API.

Note: LightOnOCR-2-1B always outputs Markdown-formatted text (including LaTeX spans for mathematical notation). The output format is embedded in the model weights and cannot be changed through text prompts. The API accepts a text field in the message content, but the model does not condition its output on it.

Features

Streaming	Reasoning	Tool calling
Not supported	Not supported	Not supported

Rate limits

Rate limits ensure fair usage and reliable access to the AI Model Hub. In addition to the contract-wide rate limits, no model-specific limits apply.