GPT-5.3 disappeared today without any in-product warning — the same issue I previously raised when GPT-5.2 was removed:
Unfortunately, today the same thing happened again — this time with GPT-5.3.
Earlier today I was still using GPT-5.3. A few hours later, it was simply gone from the model picker. No countdown, no in-product warning, no final notice in existing conversations.
What makes this especially strange is that OpenAI has already demonstrated a better approach with GPT-5.4: users were warned in advance before it was removed. So there is already a precedent for handling model retirement in a much more user-friendly way.
I understand that models are retired and products change. That is not the issue.
The issue is how the transition is communicated.
If users can build months-long conversations around a particular model, removing that model without an in-product warning breaks continuity in a way that is completely avoidable.
Release notes are documentation. They should not be the only place where users are expected to discover that a model they actively use is about to disappear.
Why wasn’t the same advance-warning approach used for GPT-5.3 that was used for GPT-5.4?
And more importantly: can in-product retirement notices become a standard part of future model migrations?