OpenAI CEO Sam Altman speaks during a media tour of the Stargate Data Center in Abilene, Texas, on September 23, 2025. Stargate is a joint venture between OpenAI, Oracle and SoftBank that is building data centers and other artificial intelligence infrastructure across the United States with promotional support from President Donald Trump.
Kyle Grillot | Bloomberg | Getty Images
OpenAI has announced that it will be retiring several models from its ChatGPT chatbot next month, including the GPT‑4o model beloved by some users.
The artificial intelligence startup launched GPT-4o in May 2024, and its warm conversational style made it very popular among some paid users. The company faced backlash in August for temporarily removing access to GPT-4o after the launch of its new model, GPT-5.
According to a post on X in August, OpenAI quickly restored access to GPT-4o for paid users, and CEO Sam Altman promised to provide “full notice” if the company decided to retire GPT-4o.
Well, that day seems to be getting closer by the minute.
“We understand that losing access to GPT‑4o will be frustrating for some users, and we did not make this decision lightly,” OpenAI said in a blog post Thursday. “It’s never easy to retire a model, but it allows us to focus on improving the model that most people are currently using.”
OpenAI said that only 0.1% of users choose to use GPT-4o every day, and the “vast majority” of people use its model GPT-5.2. The company said it feels ready to officially retire GPT-4o as it has improved the model’s personality, customization and creative ideas in recent months.
OpenAI said that apart from GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini will be deprecated from ChatGPT. The company previously announced that GPT-5 Instant and GPT-5 Thinking would also be removed from the chatbot.
OpenAI said there are no changes to the application programming interface.
Featured: OpenAI CFO Sarah Friar: There’s a massive ‘capability overhang’ happening in AI
