A large language model (LLM) is the underlying engine — a statistical model trained on enormous text datasets to predict what word should come next, over and over, until you get a coherent answer.
ChatGPT, Claude, Gemini, and Copilot are products built on top of LLMs. So an LLM is to ChatGPT what an internal combustion engine is to a Toyota Camry: the technology underneath versus the thing you actually drive.