A large language model (LLM) is an AI model trained on enormous amounts of text and other data. During training, it learns statistical patterns in language, including how words and ideas commonly relate to one another. It uses these patterns to predict and generate plausible responses.
LLMs can perform many language-related tasks, including answering questions, summarizing and translating text, generating content, analyzing information, writing computer code, and working through multistep problems. However, their responses can still be inaccurate or misleading.
Well-known AI products that use LLMs include ChatGPT, Claude, Gemini, Microsoft Copilot, and Meta AI.
You may also hear the term foundation model. This is a large, pretrained model that can be adapted for many different tasks and applications. LLMs are one type of foundation model. Some foundation models are multimodal, meaning they can process or generate more than one type of content, such as text, images, audio, or video.
When connected to web-search tools, LLM-powered products can also find current information online, summarize what they find, and provide links or citations to the sources. Because these summaries can misinterpret sources or omit important context, you should still open and evaluate the original sources.