OpenAI announces the retirement of several legacy AI models from ChatGPT.
The company sets February 13 as the final date.
GPT-4o heads the list of models facing sunset.
This marks the second attempt to retire the popular model.
In August 2025, OpenAI first tried to phase out GPT-4o after launching GPT-5.
Users protested strongly at that time.
They praised GPT-4o for superior emotional nuance and writing sophistication.
As a result, the company reinstated the model quickly.
Now, OpenAI claims newer models match or exceed those strengths.
Moreover, the user base sticking with GPT-4o has shrunk dramatically.
The company reports that only 0.1 percent of users select GPT-4o daily.
Other models retiring include GPT-5 Instant, GPT-5 Thinking, GPT-4.1, GPT-4.1 mini, and o4-mini.
This change affects only the ChatGPT interface.
The OpenAI API continues to support these legacy models unchanged.
OpenAI highlights special improvements in GPT-5.2.
Engineers fine-tuned the personality for better engagement.
They added creative ideation support and customisation tools.
Users now choose from preset base styles easily.
Additionally, markers like warmth and enthusiasm become adjustable.
Despite these upgrades, frustration spreads on social media.
Users question the 0.1 percent usage figure sharply.
One X post argues that default routing and subscription barriers skew the data.
Another user notes thousands of critical posts trending rapidly.
Interestingly, CEO Sam Altman admitted a key issue earlier this week.
During the company’s town hall, he confirmed writing quality declined in GPT-5.2.
The team shifted heavy focus to coding, reasoning, and mathematics.
Therefore, many longtime users feel disappointed.
They argue the retirement pushes them away from ChatGPT.
Overall, this decision sparks debate about progress versus preference.
OpenAI insists the transition benefits everyone.
However, loyal GPT-4o fans remain vocal and resistant.