OpenAI to Retire ChatGPT Models: What You Need to Know
â–Ľ Summary
– OpenAI is permanently retiring the GPT-4o model from ChatGPT’s user interface on February 13, along with GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini.
– The company previously removed GPT-4o last August but restored it after significant user backlash over its perceived warmer, friendlier personality compared to newer models.
– OpenAI’s CEO, Sam Altman, promised better notice for future retirements and more user customization, leading to the current availability of multiple model personalities.
– The company justifies the retirement by stating that only 0.1% of users interact with GPT-4o daily, though the models will remain available via the API for developers.
– This incident sparked broader discussions about user emotional attachment to AI models and the ethics of model deprecation.
OpenAI is preparing to remove several older ChatGPT models from its platform, a move that will directly impact how users interact with the popular AI assistant. The company periodically phases out legacy models as newer, more advanced versions become the standard. This latest retirement wave, scheduled for February 13, includes specific models that a small fraction of the user base still relies on.
The most notable departure is GPT-4o, a model whose initial removal last August sparked significant user protest. Many found its conversational tone warmer and more personable compared to the more direct responses of its successor, GPT-5. The backlash was so pronounced that OpenAI quickly reversed its decision, reinstating access while promising ample future notice before any final deprecation. This episode highlighted a surprising depth of user attachment to specific AI personalities, prompting internal discussions about ethics and the need for greater user control.
Following that event, CEO Sam Altman committed to enhancing customization options. The company has since delivered on that promise, introducing a system where users can select from several distinct personality presets or define their own through custom instructions. This shift toward personalization aims to balance user preference with the technical and economic realities of maintaining multiple model versions.
Now, with user adoption metrics showing that only a tiny segment actively uses the older models, OpenAI is moving forward with the retirement. Alongside GPT-4o, the list of models being removed from the ChatGPT interface includes GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini. It is crucial for developers to note that these models will remain accessible via the API for the time being, with no immediate changes planned for that developer-facing service. This ensures that applications and integrations built on these models can continue to function without interruption.
The company’s approach reflects a balancing act: streamlining the consumer product experience by focusing resources on its most capable and widely used models, while still supporting the developer ecosystem that depends on API stability. For everyday ChatGPT users, the transition underscores the evolving nature of the tool and the increasing emphasis on tailoring the AI’s interaction style to individual preference.
(Source: 9to5 Mac)
