# What is Vision-language-action models? Definition and meaning

A neural network that takes camera input and a natural-language instruction and outputs low-level actions for a robot.

## Definition

**Vision-language-action models**: A neural network that takes camera input and a natural-language instruction and outputs low-level actions for a robot.

In short: See, understand, then move.

## When it is used

General-purpose robot manipulation, humanoid control, household and warehouse tasks, and robotics research.

## Where it fits

Vision-language-action models belongs to **Frontier topics** in the [Deep learning](/topics/deep-learning) topic. These are the other terms in the same group:

- **[World models](/glossary/world-models)**: A learned model that predicts how an environment will change, given its current state and an action, so an agent can plan or train inside it.

## Learn more

The full guide to **[Vision-language-action models](/tools/vision-language-action-models)** explains how it works, when to use it and how it compares with the alternatives, with examples. The [glossary](/glossary) lists every term on the site from A to Z.

Previous term: [Vespa](/glossary/vespa) · Next term: [ViT](/glossary/vit)
