The maximum number of tokens to generate. Ensure that the sum of the input tokens and max_tokens does not exceed the model's context window.

The maximum number of tokens to generate. Ensure that the total number of input tokens does not exceed the context window of the model. Since some services are still being updated, it is recommended not to set max_tokens to the upper limit of the window; reserve a buffer of about 10k tokens for input and system overhead. For more details, please refer to Models。