OpenAI has officially announced a major model iteration plan: starting next month, several older large models, including GPT-4o, will be discontinued from ChatGPT. This decision marks the end of the historical mission of the once highly praised GPT-4o, which is now entering the technical museum.

Since its release in May 2024, GPT-4o has won a loyal following with its friendly conversational style and excellent multimodal capabilities. However, with technological advancements, OpenAI has fully shifted its focus to the new generation of flagship models.

Survival of the fittest: GPT-5.2 has become the absolute choice for users

OpenAI's data shows that currently, only 0.1% of users are still using GPT-4o. In contrast, the majority of users have already migrated to the more powerful and logically sound GPT-5.2 model. The official stated that in recent months, the company has made significant progress in personalization, customization, and creative thinking capabilities of the model, and it is now the right time to discontinue the old model.

Cleaning list: Several "mini" and reasoning models are also leaving at the same time

The scope of this "clean-up" goes beyond GPT-4o. According to the announcement, GPT-4.1, GPT-4.1mini, and OpenAI o4-mini will also be removed from the ChatGPT service list. Additionally, the previously released GPT-5Instant and GPT-5Thinking are also on the list of models to be discontinued.

Previously, when OpenAI attempted to disable access to GPT-4o, it was temporarily restored due to user dissatisfaction. CEO Sam Altman at the time promised to notify users in advance if they were to discontinue it. Now, with the full rollout of the GPT-5 series models, OpenAI has finally made up its mind to streamline the model library and concentrate computing resources to provide users with a more unified and advanced AI interaction experience.

Would you like to know more:

  • What are the new features of the latest GPT-5.2-Codex API opened by OpenAI for developers?

  • What are the three versions of the rumored upcoming GPT-5.1 targeting which application scenarios?

  • Why did industry figure Yann LeCun publicly criticize the mathematical capabilities of GPT-5?