New member
- Messages
- 5
I've been throwing this question around a few places and haven't gotten a straight answer. Now that ChatGPT has project folders with custom instruction sets and upload abilities, I've been porting some of my instructions for personal custom GPT's to projects. As of now you can't select what model will run the custom GPT, but you can select different models per project conversation (and even change mid conversation). If you don't need to make it public and you don't want to share it, and you also don't need the function calling, why not use projects instead? 
