OpenAI Retires Beloved GPT-4o Despite User Protests

OpenAI is pulling the plug on GPT-4o, the conversational AI model that some users swear by for its warm, human-like interactions. Despite passionate user advocacy and previous protests, the numbers tell a different story: only 0.1% of users choose GPT-4o daily.

The announcement, made Thursday, marks the end of an era for a model that launched in May 2024 to considerable fanfare. GPT-4o's distinctive conversational warmth made it a favorite among a dedicated subset of paid users, leading to significant backlash when OpenAI briefly removed access last August.

The Promise Kept

CEO Sam Altman had promised "plenty of notice" before retiring popular models, a commitment made after the August controversy forced the company to quickly restore GPT-4o access. True to his word, OpenAI is now providing that advance warning.

"We know that losing access to GPT‑4o will feel frustrating for some users, and we didn't make this decision lightly," the company stated in its blog post. "Retiring models is never easy, but it allows us to focus on improving the models most people use today."

The retirement list includes GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini, alongside the previously announced GPT-5 Instant and GPT-5 Thinking. The company emphasized that its API services remain unchanged.

The Tyranny of Small Numbers

The 0.1% usage figure might seem negligible, but context matters. With ChatGPT boasting over 300 million monthly active users, that translates to potentially 300,000 people who prefer GPT-4o's conversational style. That's roughly the population of a mid-sized city.

OpenAI justifies the decision by pointing to the "vast majority" of users who've migrated to GPT-5.2, along with recent improvements in model personality, customization, and creative capabilities. The company argues these enhancements make GPT-4o's retirement feasible.

Yet this raises uncomfortable questions about how tech companies define "majority" and whether statistical minorities deserve consideration in product decisions.

The Efficiency vs. Choice Dilemma

This move reflects a broader challenge facing AI companies: balancing operational efficiency with user preference diversity. Maintaining multiple models requires significant computational resources and development attention. From a business perspective, focusing on the most popular option makes sense.

But there's a philosophical dimension too. GPT-4o's fans aren't just attached to outdated technology—they're advocating for a specific type of AI interaction that feels more natural and engaging to them. The retirement suggests that in AI development, popularity often trumps preference diversity.

The timing is particularly notable given the competitive landscape. As AI companies race to deploy more capable models, they're simultaneously narrowing the range of interaction styles available to users.