I have seen several posts that indicate 4.5 is rolling out. A gpt+ user can type in “what LLM architecture does your api use” and some users are getting the answer 4.5. Noticeable improvements in speed and quality. Let me know if any of you have that experience please.