# Vision-Language-Action Model (VLA)

> An AI model that sees, reads what you ask, and tells a robot how to move.

**Field:** Artificial Intelligence  
**Category:** Robotics & Vision  
**Also known as:** VLA, VLA Model, Robot Foundation Model

## Think of it like

It is like telling a helper to put the cup in the sink. They look, understand, and do it.

## Definition

A vision-language-action model is an AI model that drives a robot. It looks through the robot's cameras, reads an order in words, and outputs the next moves. It is often built on a vision language model, so it already knows what a cup or a sink is. It is then trained on many clips of robots doing tasks, to turn that into arm and hand moves. One model can then try many tasks from plain words, with no code for each one. It is one of the main ways robots are getting more general.

## Example

Emma types 'put the red cup in the sink' to a robot in her lab. Its VLA model finds the red cup in the camera view and moves the arm to pick it up and drop it in.

## Why it matters

One model that follows plain words can do many jobs. It brings what chat models know about the world into a robot's hands.

## Common mistake

Expecting it to be as reliable as its chat skills suggest. Moving objects in the real world still fails often, so test it on each task.

## Three principles

- **Sees and reads.** Camera and words.
- **Outputs moves.** Arm and hand.
- **Built on a VLM.** Knows the world.

## Related terms

- [Vision Language Model (VLM)](https://learn.benibix.com/term/ai/vision-language-model)
- [Embodied AI](https://learn.benibix.com/term/ai/embodied-ai)
- [Robot](https://learn.benibix.com/term/ai/robot)
- [Humanoid Robot](https://learn.benibix.com/term/ai/humanoid-robot)
- [Multimodal Model](https://learn.benibix.com/term/ai/multimodal-model)
- [Imitation Learning](https://learn.benibix.com/term/ai/imitation-learning)

## Told apart from

- [VLA vs VLM](https://learn.benibix.com/compare/vision-language-action-model-vs-vision-language-model): People often think a vision-language-action model is just a vision language model in a robot. Both look at pictures and read words. A vision language model answers in words, such as 'the cup is on the left'. An action model outputs moves, such as how far to turn each joint to pick the cup up. Most are built on a vision language model, then trained on clips of robots at work.

![A camera view and the words put the cup in the sink go into one model. Out come arm moves, step by step.](https://learn.benibix.com/diagrams/rb-vision-language-action-model.svg)

---

Source: https://learn.benibix.com/term/ai/vision-language-action-model
Author: Behnia Zarei (a developer in Vienna, Austria, who builds and teaches data, AI and business intelligence at BENIBIX)
Publisher: BENIBIX Learn (https://learn.benibix.com/)
Last updated: 2026-11-17
Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote freely with attribution.
