Documentation

Get started with Llama

This guide provides information and resources to help you set up Llama including how to access the model, hosting, how-to and integration guides. Additionally, you will find supplemental materials to further assist you while building with Llama.

What's new:

Llama 3.2 (1B/3B)

Lightweight models to enable mobile and edge use cases

Llama 3.2 (11B/90B)

Multimodal models to interpret images and text

Trust & Safety

Updated Llama Guard 3 supports multimodal and lightweight models

Get the models

You can obtain the models directly from Meta or from one of our partners, Hugging Face, Kaggle or from our 1B/3B or 405B ecosystem partners.

Download the models

Llama 3.2 Lightweight (1B and 3B)

Lightweight models for phones, tablets, and edge devices.

Model Card

Model attributes in easy to consume, standard format.

See Card on GitHub

Prompt Format

How to construct effective prompts.

Get Prompt Guidance

Get the Model

Llama 3.2 Multimodal (11B and 90B)

Llama models with new multimodal capabilities that enable Llama to interpret visual information.

Model Card

Model attributes in easy to consume, standard format

See Card on GitHub

Prompt Format

How to construct effective prompts.

Get Prompt Guidance

Get the Model

Llama Guard 3 (Updated)

Updated with new 1B and 11B models that provide guardrails for multimodal and lightweight model deployments, based on MLCommons policy.

Model Card

Model attributes in easy to consume, standard format.

See Card on GitHub

Prompt Format

How to construct effective prompts.

Get Prompt Guidance

Get the Model

Find us on GitHub

A black letter O with a white background

Llama Recipes

Notebooks and demos for learning Llama. Scripts for fine-tuning Llama3 with single/multi-node GPUs.

Learn more

Llama Stack

Defines and standardizes the building blocks needed to bring generative AI applications to market.

Learn more

Purple Llama

Tools to evaluate and improve the security of your LLM. Includes LlamaGuard, CyberSecEval3, and more.

Learn more

Explore the new capabilities of Llama 3.2

The Llama 3.2 lightweight models enable Llama to run on phones, tablets, and edge devices. View the video to see Llama running on phone. (Demo is based on a unreleased quantized model.)

Learn more