Support for retrieving the maximum number of generatable tokens per request.

[Valkryst]

As far as I have seen, and please correct me if I'm wrong, the SDK doesn't have a way to programmatically retrieve the maximum number of tokens that can be generated in a single request.

I'd like to see these values added to the ChatModel enum, or another appropriate area.

Here're a few examples, just in-case I'm using the wrong terminology:

• gpt-3.5-turbo -> Max of 4096
• gpt-4 -> Max of 8192
• gpt-4-32k -> Max of 32,768

These values were pulled from an old project of mine, which in-turn pulled them from an old copy of the OpenAI docs.