Meta's foundation-model and applied-AI research division. Releases the Llama family of open-weight large language models, which underpin a large portion of the open-source LLM ecosystem. Also publishes research on multimodal, agents, and recommendation.
