Active
Meta AI
Meta's AI assistant and the Llama family of downloadable models.
- Meta AI assistant for questions and image creation
- Llama model weights for custom deployments
- Separate terms for the assistant and each model release
Meta AI includes a finished assistant, while Llama is a family of models for building applications. Llama weights are available from Meta and distribution partners. These are different choices: using the assistant does not provide a private model deployment or a general-purpose developer API account.
Choosing an access route
A downloaded model requires suitable compute and compliance with its release license. Cloud-hosted Llama deployments follow the chosen host's pricing. Assistant availability and features depend on the supported account, product, and region.
Sources
From the archive
- SEA-LION v3.5 adds reasoning and a choice of response modes 2025-04-15
- Llama 4 Scout and Maverick add native image understanding 2025-04-05
- Cerebras opens its inference service to developers 2024-08-27
- Ollama adds tool calling to local model requests 2024-07-25
- Llama 3.1 adds a 405B model and longer multilingual context 2024-07-23
- Together adds Turbo and Lite endpoints for Llama 3 2024-07-18