I've written a UserScript which allows me to better control which model to use, so I thought I'd share it.

I started writing the UserScript a couple of months back, because I'd rather get faster GPT3.5 response, than getting limited with 4o every now and then.

It's hacky at best, but it does the job. In short, choose either:

- Auto, where OpenAI would decide what's best

- GPT3.5, available until OpenAI actually sunsets the model

- GPT 4o mini, so you can save your 4o calls for more complex queries

https://github.com/altbdoor/userscripts/raw/master/force-gpt...

You will need a UserScript browser extension, and then install the extension above:

- https://violentmonkey.github.io/ - https://www.tampermonkey.net/