Model

A large language model (LLM) learned patterns from a huge amount of text and predicts, piece by piece, the most likely continuation of what it receives. That is what lets it answer questions, write code or summarize documents.

The same model can show up in many products. A chat in the browser, an agent tool and an API are different paths to the same model, with different rules and tools around it.

A model is not a product

The product is what surrounds the model: interface, history, tools, permissions. That is why the same model can feel more capable in one place than another. That surrounding layer has a name: the harness.