LLM vs Agent vs Model: Plain-English Hierarchy for Developers
An LLM is a text predictor: you send words in, it predicts the next token until the reply is done. An agent wraps that same model with tools and a plan so it can search, write files, or call APIs. A subagent is a helper agent spun up by another agent for one slice of a bigger job. Same stack,…