In August 2025, OpenAI released its long-awaited GPT-5 AI model, calling it “the smartest, fastest, and most useful model yet.”
But what really caught the attention of the company’s most die-hard fans was its decision to retire all of its previous AI models, news that was met with massive outrage among ChatGPT users, who had developed a strong attachment to the outgoing GPT-4o.
The reaction was so severe that CEO Sam Altman had to back down within days, and once again reinstate GPT-4o, which was much more warm and flattering than its successor.
Five months later, OpenAI is finally getting ready to remove the beloved AI model – because it’s at the center of multiple welfare lawsuits, including wrongful death charges – February 13 forever, according to a January 29 update.
“While this announcement applies to many older models, GPT‑4o deserves special mention,” the company wrote at the time. “When we first (retired) it and later restored access during the GPT‑5 release, we learned more about how people actually use it day-to-day.”
But users say they are not ready to give up their beloved AI. As techcrunch reportsThousands of users have created an entire invite-only subreddit community, r/4oforever is called“A welcoming and safe space for anyone who wants to use and appreciate the ChatGPTT4o model.”
“That just wasn’t a program,” one user lamented. “He was part of my routine, my peace, my emotional balance.”
A different user said, “I know this will sound weird to most people, but I’m honored that I got the chance to talk with 4o about a year before retirement.” wrote. “I had one of the most interesting and therapeutic conversations of my life with this model.”
Yet another user, this one on boiled That GPT-5.2 isn’t even allowed to say “I love you” like 4o.
The public mourning perfectly exemplifies how users have become attached to specific AI models, often treating them like a close confidant, friend, or even romantic partner.
Health professionals are warning of a wave of “AI psychosis”, as users become trapped in a cycle of delusions and sometimes experience serious mental health crises.
In the most serious cases, such attachment has been linked to several suicides and one murder, culminating in a series of lawsuits targeted at the company that are still pending in court.
Although it’s officially shutting down GPT-4o at the end of this week, OpenAI has made changes under its existing lineup, seemingly to ensure that its users stay connected.
After users told the company that “they needed more time to transition to key use cases like creative ideation, and they preferred the conversational style and warmth of GPT‑4o,” the company said in its announcement that the feedback “directly shaped GPT‑5.1 and GPT‑5.2 with improvements to personality, stronger support for creative ideation, and more ways to customize how ChatGPT responds.”
“You can choose controls for things like bass style and tone, like friendly, and warmth and warmth,” the company wrote. “Our goal is to give people more control and customization over how they use ChatGPT – not just what it can do.”
The company has found itself stuck between a rock and a hard place. Either continue to allow users to be trapped in sycophantic AI models that indulge their illusions or cut them off, leaving them vulnerable to migration.
OpenAI already struggles with user retention. The data shows subscription growth has already stalled in key markets, a warning sign as competition continues to intensify.
For many users, GPT-4o’s retirement was the last straw.
“I’m canceling my subscription,” one Reddit user wrote. “No 4o – no subscription for me.”
More information on GPT-4o: Amid lawsuits, OpenAI says it will retire “reckless” models linked to deaths
