Menu
Home
People
Places
Arts
History
Plants & Animals
Science
Life & Culture
Technology
Reference.org
Llama (language model)
open-in-new
<p>Llama (Large Language Model Meta AI, formerly stylized as LLaMA) is a family of <a href="/facts/Large_language_model/WnogWVJY">large language models</a> (LLMs) released by <a href="/facts/Meta_AI/raPFo7MW">Meta AI</a> starting in February 2023. The latest version is Llama 4, released in April 2025. </p><p>Llama models come in different sizes, ranging from 1 billion to 2 trillion parameters. Initially only a <a href="/facts/Foundation_model/67pefJcT">foundation model</a>, starting with Llama 2, Meta AI released instruction <a href="/facts/Fine-tuning_(deep_learning)/9eRVyFNy">fine-tuned</a> versions alongside foundation models. </p><p>Model weights for the first version of Llama were only available to researchers on a case-by-case basis, under a non-commercial license. Unauthorized copies of the first model were shared via <a href="/facts/BitTorrent/HP9ChWWW">BitTorrent</a>. Subsequent versions of Llama were made accessible outside academia and released under licenses that permitted some commercial use. </p><p>Alongside the release of Llama 3, <a href="/facts/Meta_Platforms/ylApt16y">Meta</a> added <a href="/facts/Virtual_assistant/BuouwL2U">virtual assistant</a> features to <a href="/facts/Facebook/0IhXFR6u">Facebook</a> and <a href="/facts/WhatsApp/VMlSeFb6">WhatsApp</a> in select regions, and a standalone website. Both services use a Llama 3 model. </p>