Did you start using GPT-4.5-turbo-preview in production?
Is there something that really prevents us from starting really using it?
From what I heard, Microsoft and other big partners use it in production since weeks or even months, so I guess it’s already kind of working and stable. It’s such a pricing and speed game changer that it’s heart-breaking to keep not using it!
You can review the models page of documentation to find which models are disclaimed “not yet ready for production use”.
Shared “turbo” limits:
gpt-4-turbo-preview (points to gpt-4-0125-preview)
gpt-4-0125-preview (newest model)
gpt-4-1106-preview - will never be “ready”, as it has some flaws.
All of these will now have higher token-per-minute rates than full non-turbo’d GPT-4, but still have a token-per-day limitation if not in a higher tier (resulting from previous API payments to OpenAI.)