The generated output from a Large Language Model (LLM) in response to a given prompt. This term originates from the way prompts are usually structured, requiring the model to complete a sentence or thought.