Large language model

An LLM predicts text: given what came before, it produces the most fitting continuation. At large scale, that ability becomes answering questions, summarizing, writing and following instructions. GPT, Claude, Gemini and Grok are families of LLMs. An LLM on its own is not an agent. It has no access to your orders, cannot take actions and only knows what was in its training data or the conversation. An AI agent wraps an LLM with tools, data, memory and rules, so it can look things up, act and check its results. Models differ in speed, cost and how well they handle long, multi-step work. Many agent products let you choose the model, or use different models for different parts of the job.