AI Orbit A2Z Logo

🇺🇸 United States · USA

Meta (Llama) models explained

Meta releases the Llama family as open-weight models that anyone can download under Meta’s licence, and uses them in the Meta AI assistant across WhatsApp, Instagram and Facebook.

All model families

Open-weight chat

Llama (large)

Best for: Self-hosted assistants on servers

Large versions need data-centre GPUs; available through many cloud providers.

Open weightsAPI

Open-weight chat

Llama (small)

Best for: Laptops, phones and edge devices

Compact versions that run locally with Ollama or LM Studio.

Open weights

Text + image

Llama vision / multimodal

Best for: Understanding images and documents

Versions that accept images as input.

Open weights

Assistant

Meta AI

Best for: Chat inside WhatsApp, Instagram, Messenger

Meta’s consumer assistant built on Llama.

App

Model names and versions change often. Check the official documentation for the newest releases.

⚡ Which model should you use?

  • Try Meta AI inside WhatsApp for casual use.
  • Developers: pick the smallest Llama that meets your quality bar and run it locally or in the cloud.

Data, privacy and your region

US providers offer business and enterprise plans with data-processing agreements, and some offer regional data residency (including in the EU/UK). Check the plan’s data-use and retention settings before sharing sensitive information. Check Meta’s licence terms for commercial use.

Frequently asked questions

Is Llama open source?

Llama models are open-weight — you can download and run them — under Meta’s own licence, which has some conditions for very large companies and certain uses.

Step-by-step guideFull tool review

More AI companies