ChatGPT wasn't even how this technology was meant to be used. They simply didn't know how to package it. You can tell from all these models have structured response output. It was probably made for structured response but became useful in other ways. Using ChatGPT is equivalent of driving a golf cart on the highway. Most people are not smart enough to use it. ChatGPT doesn't even tell you how far up into the context it will still consider the text, so you're always at risk it might forget previous information when you go beyond the first response, there is simply no visibility or guarantee. I call it, fog-of-context. GPT5.4 is $15. GPT5.4-pro is $180, how will the next model be priced? What can pro model do that you can't do with any other model and why did they remove cached pricing? Don't get me started about the input. We shouldn't even be paying for that. They should be happy some of us continue to use OpenAI despite other's opinions. Sam Altman should be kissing us, not them.
