OpenAI has announced that several older ChatGPT models will be removed from the platform on February 13, 2026, as the company continues to consolidate around its newer GPT-5 lineup. The models being retired from ChatGPT include GPT-4o, GPT-4.1, GPT-4.1 mini, o4-mini, along with earlier GPT-5 Instant and Thinking variants. OpenAI clarified that this change applies only to ChatGPT itself, with no immediate impact on API users.
Among the models being phased out, GPT-4o stands out due to its unique history. Although it was previously marked for deprecation, OpenAI later reinstated it after feedback from Plus and Pro subscribers who relied on its conversational tone and creative flexibility. Many users felt GPT-4o offered a warmer, more natural interaction style that was especially useful for brainstorming and creative work.
That feedback, OpenAI says, played a key role in shaping GPT-5.1 and GPT-5.2. The newer models introduce refined personality controls, improved creative output, and expanded customization options, allowing users to adjust tone, warmth, and enthusiasm to better suit their preferences.
With these improvements now fully implemented, OpenAI believes the transition period has effectively ended. Usage data shows that GPT-5.2 now accounts for virtually all daily interactions, with GPT-4o representing just 0.1% of active use. As a result, OpenAI considers the retirement of older models a natural step rather than a disruptive one.
Beyond model consolidation, the company also reiterated its broader goals for ChatGPT’s future. Upcoming updates are expected to further enhance creativity and personality, while also reducing overly cautious behavior and unnecessary refusals — areas users have repeatedly highlighted as pain points.
