A lot of complains recently are referring to the MAX models costing extra money, and standard models being ‘dumbed down’ because of shorter context windows.
As professional developers, we value visibility and control. A lot of these misunderstandings would be cleared if we were able to see exactly how much context we are consuming and what we are and are not sending in our requests. I want to be able to decide by myself how much context I want to use, and how much it is going to cost me.
How about a context size slider, with associated costs (e.g. number of premium requests it will take)? How about a token counter for the currently active chat?
Give us visibility and control, and the option to consume more premium requests, rather than additional pricing layers that make us feel like we’re being ripped off with dumber model versions.