Raw LLM Responses

New in terms of 'they didn't release any updated versions of the old models, even though they kept working on them, so they could use them for the gpt5 amalgam and get better benchmarks'. Not like it is some groundbreaking architectural change that drastically upgraded capabilities. Slight updates on a portion of the niche tools with the added overhead of extra calls is more than enough. Faster? Can't tell. Would be interesting to be able to determine, if only they didn't have extra delays since launch to spread out the frequency of calls. At this point any 'faster' models on the ChatGPT website could just be decreasing the delay before showing the text. I'd bet it claims it is using the exact model you asked for, just like it will claim its an astronaut or a plumber if you ask it. Too bad the other tools were taken away and everything has to be behind smoke and mirrors.