What is a large language model (LLM)?

A large language model is the technology behind chat assistants and many AI writing features. It is trained on a very large collection of text, and in the process learns patterns in how words, facts, and ideas tend to fit together. When given a prompt, it produces a response one small piece of text, called a token, at a time, each time choosing a likely next piece based on everything before it.

In everyday work, an LLM is what drafts an email from a few bullet points, summarizes a long report, rewrites a paragraph in a friendlier tone, pulls names and dates out of a contract, or answers questions about a document that has been pasted in. The same model can do all of these, because each is a variation on reading and producing text.

An LLM does not look facts up the way a search engine does unless it is connected to search or to documents. It generates text that sounds right based on patterns, which is why it can state false things confidently, a problem known as hallucination. Its built-in knowledge also stops at the point its training data was collected, so it may not know about recent events.

Names like GPT, Claude, and Gemini refer to families of LLMs made by different companies, and chat products are built on top of them. The model shapes the quality of reasoning and writing, while the product around it adds features such as file uploads, web search, and connections to other apps. Output still needs checking, especially facts, figures, and anything sent to customers.

An example

A manager pastes meeting notes into an LLM-based assistant and asks for a short summary with action items and owners.