OpenAI occasionally phases out older models of ChatGPT as users adapt to the latest ones. Recently, the company announced the retirement of several models, including GPT-4o, on February 13.
When OpenAI launched GPT-5 last August, it unexpectedly removed GPT-4o from the model selection. This decision upset many users who preferred GPT-4o’s friendly tone compared to GPT-5’s more direct style. The backlash led OpenAI to restore GPT-4o and sparked discussions about the emotional connections people form with AI.
OpenAI’s CEO, Sam Altman, acknowledged the situation, stating that they would notify users well in advance before any future retirements. He also mentioned the need for greater customization options for users, which they are now addressing. Users can now choose from seven different personalities for ChatGPT, along with setting custom instructions to tailor their experience.
Despite this, OpenAI has decided to retire GPT-4o again, citing that only 0.1% of users interacted with it daily. Other retiring models include GPT-4.1, GPT-4.1 mini, and OpenAI’s o4-mini. However, these models will remain available through the API for developers.
Interestingly, the rise in AI usage has been accompanied by evolving user expectations. A recent survey revealed that over 60% of users prefer models that can adapt their tone based on context. This trend highlights a growing demand for personalized AI interactions, which OpenAI seems to be listening to.
For those interested in understanding more about these model changes, you can read OpenAI’s announcement here.
