What’s the new OpenAI GPT-4o situation?
GPT-4o remains “retired” for some users
Coverage indicates that OpenAI replaced GPT-4o, and some users still haven’t accepted the change. Instead, a subset of fans has organized a public celebration—shown as a video billboard in Times Square—marking a “birthday” for the model.
The point of the tribute is straightforward: these users want GPT-4o back, treating the model as something more than just another version number. That matters because model transitions affect real workflows—especially for users and developers who have built expectations around a particular model’s behavior.
At the same time, the underlying premise of the story is that GPT-4o is no longer the active default for at least some parts of OpenAI’s offerings after a replacement. Even so, user sentiment can remain sticky when there’s a perceived change in quality, style, or capability from one model to another.
This is also a window into a recurring dynamic in the AI market: models aren’t only products; they become “interfaces” people learn and optimize for. When a provider swaps which model powers a feature, users may experience friction, and communities may respond quickly—through petitions, unofficial workarounds, or, in this case, a public campaign.
What to watch next
- Whether OpenAI offers a way to access GPT-4o-like behavior
- Whether community pressure influences product decisions or pricing
- Whether the “retired” status changes for specific user groups or plans
The coverage doesn’t specify which exact GPT-4o access is or isn’t available, so the only confirmed part is that some users are still lobbying for the model’s return after it was replaced.