Introduction to Meta’s Llama 4 Herd

The rollout of Meta’s Llama 4 herd on Azure AI Foundry has begun. This development is exciting due to the models’ ability to integrate text and vision tokens seamlessly, which sets a new benchmark for applications that require agency.

Key Features of Llama 4 Models

• Llama 4 Scout: Notably, Llama 4 Scout can operate within a single H100 GPU, supporting a near-infinite 10m token context window. This capability makes it particularly suitable for tasks such as multi-document summarization and reasoning over extensive codebases.
• Llama 4 Maverick: With 17b active parameters, Llama 4 Maverick excels in understanding both images and text, supporting up to 12 languages. It is ideal for creating sophisticated applications that can bridge language barriers or for interactive scenarios like customer support.

Further Information and Acknowledgments

For more details on Llama 4, visit https://www.linkedin.com/signup/cold-join?session_redirect=https%3A%2F%2Fwww.linkedin.com%2Ffeed%2Fhashtag%2Fllama&trk=public_post_reshare-text or explore the Microsoft Azure showcase. Additional information can be found at https://lnkd.in/g_nt766N. Special thanks to Connor Hayes, the Meta team, and the members of our AI Platform team for their contributions.