The Llama series of large language models (LLMs) was developed by Meta AI, with the initial version released in February 2023. Designed as autoregressive decoder-only Transformers, Llama models work by predicting the next word in a sequence, thus generating text. The latest iteration, Llama 4, launched in April 2025, introduces native multimodality, enabling the processing of both text and image inputs. These models come in various sizes, indicated by the number of parameters, with the aim of providing accessibility for diverse hardware capabilities. Trained on vast datasets of publicly available text, and now incorporating image and video data in Llama 4, these models have demonstrated strong performance in various natural language processing tasks and are intended to support research and development in the AI community.