Large Language Model (LLM)
Also known as · LLM
An AI model trained on vast amounts of text to predict and generate language.
A large language model is an AI system trained on enormous quantities of text to do one deceptively simple thing: predict the next token (a word or word-fragment) given everything before it. Repeat that prediction over and over and the model produces fluent paragraphs, answers, and code.
What makes them "large" is scale — billions of internal parameters tuned across trillions of words. At that scale, abilities that were never explicitly programmed start to emerge: following instructions, translating, summarizing, reasoning step by step. The model isn't looking anything up; it's drawing on statistical patterns compressed from its training data.
Modern LLMs like Claude, GPT, and Gemini are general-purpose: the same model can write an email, debug code, or explain a contract. That generality is why they've become a platform technology rather than a single-purpose tool.