OpenAI to retire GPT‑4o and legacy models from ChatGPT on 13 February
Company cites low usage and transition to GPT‑5.2 as reason for removal
OpenAI will remove GPT‑4o, GPT‑4.1, GPT‑4.1 mini and OpenAI o4‑mini from ChatGPT on 13 February, the company said in a statement.
The company initially deprecated GPT‑4o but restored it during the release of GPT‑5 following feedback from users who preferred its conversational tone and relied on it for creative ideation.
OpenAI said the temporary return gave users time to adapt their workflows and that the feedback informed improvements in GPT‑5.1 and GPT‑5.2, including enhanced personality, better creative support, and new tools for setting response tone and warmth.
OpenAI said GPT‑5.2 now accounts for the vast majority of usage, with only 0.1% of users selecting GPT‑4o daily.
Related reading
- OpenAI develops internal AI agent to accelerate data-driven decision making
- OpenAI limits agent access to unverified URLs to curb unauthorised data transfer
- Microsoft unveils UniRG to improve accuracy in AI-generated radiology reports
The company added that it is continuing work on refining personality and creativity, reducing unnecessary refusals, and developing a version of ChatGPT intended for users over 18. Age prediction features have already been rolled out in most markets for under-18 users.
There are no changes to the API at this time. OpenAI said retiring legacy models allows the team to focus on improving the models most users rely on.
The Recap
- OpenAI will retire four legacy ChatGPT models on 13 February.
- Just 0.1% of users used GPT‑4o each day.
- The API will have no changes at this time.
