Allow use of gpt-4-32k model

[archvalmiki]

Currently, specifying that in model doesn't work. Allow this option to be used in conjunction with max tokens to be set to 32,768.

[yigitkonur]

You could try going and editing the plugin's own code to remove the places where the max_tokens parameter is. This parameter is no longer mandatory - if you don't need to limit the AI, I think not using max_tokens will solve this problem.