Vision-language model (VLM)

A vision-language model reads images, and often video, together with text, so you can ask questions about a picture, a screenshot or a document page.

  • Models and architecture
  • Updated

A VLM pairs a language model with a vision encoder that turns an image into tokens the language model can read. The output is text, so a VLM can describe a photo, read a chart, answer questions about a screenshot or turn a page into Markdown. The term is narrower than multimodal model, which also covers audio and the generation of images or speech.

Many 2026 releases are VLMs by default, for example Qwen3.5, Gemma 4 and Mistral Large 4. The OCR models on this site are small VLMs trained for one job, reading documents. Check the input column on each model page, because image input does not mean video or audio input.

Related terms

  • Multimodal model A multimodal model can handle more than one type of data, such as text, images, audio or video, as input, as output, or both.
  • OCR and document parsing OCR turns an image of text into text; document parsing goes further and rebuilds tables, formulas, charts and reading order, usually as Markdown.
  • Token A token is a small chunk of text, often a word or part of a word, that a language model reads and writes; models measure length and speed in tokens.

Keep reading

All terms