The maximum number of tokens (words or word pieces) a model can generate in a single response, controlling the length of its output.
Code generation, debugging, explanation, and refactoring