Many people are observing ChatGPT-4 claim to be "gpt-4.5-turbo":

https://chat.openai.com/share/e0458ead-2b28-4344-941f-9adfaa13dfe0

I'm also seeing people say on Twitter that GPT-4 has gotten better in coding ability the last day or so.

It could be part of a rolling release or A/B test. Seems like an odd mistake for it to make, especially happening just after false leaks of /Mira/will-gpt45-be-announced-on-thursday . Could be real, could be an OpenAI psyop, or could be a coincidence.

When GPT-4.5 is officially announced, will OpenAI mention in their announcement that these are genuine GPT 4.5 model interactions?

Resolves YES if OpenAI confirms that they have been slowly rolling out GPT 4.5 prior to the announcement. Either in announcement or by recording the model id in a traceable location(ChatGPT itself isn't reliable, but its system prompt would be).

Resolves NO if it is GPT 4.5 but it isn't confirmed anywhere.

Resolves NO if GPT 4.5 isn't released by end of year.