I want to know whether the high max-token models with $ per request actually deliver better results for money compared to regular models when it comes to fast queries with a large context.
I’m already familiar with all the theoretical “blah-blah” about the differences — what I really want is feedback from people who have actually used them: do you actually see a difference or not?