ah this is the same as the GPT4 question - we're currently building the ability to support other models