Switch models during chat

[gwgrant7]

Would it be possible to update the model during a chat for example start with 3.5 and the switch to gpt4 for a more precise response.

This option would help with the message limitations by allowing you to only use gpt4 and plugins when required.