OpenAI will remove GPT-4o and several older ChatGPT models from the app on February 13, 2026, prompting user backlash.
• OpenAI will remove GPT-4o, GPT-4.1, GPT-4.1 mini, and o4-mini from ChatGPT on Feb 13, 2026 • OpenAI says only 0.1% of users still choose GPT-4o daily; API access remains unchanged for now • Users reacted with frustration and sadness; company points to GPT‑5.1/5.2 personality controls as replacements
OpenAI/product team: Retiring underused models lets the company focus engineering and moderation efforts on newer GPT‑5 variants and offers users new personality and customization options. Dedicated users/communities: Many feel emotionally attached to GPT‑4o’s tone and style and view the short notice as insensitive and disruptive to personal and creative workflows. Developers/business users: Welcome that API access is unchanged for now, but some will watch timelines closely to plan migrations and ensure feature parity with newer models.
OpenAI announced it will retire several ChatGPT models — specifically GPT‑4o, GPT‑4.1, GPT‑4.1 mini and OpenAI o4‑mini — from the ChatGPT model picker on February 13, 2026, while saying those models will remain available via the API for now. The company told users that the vast majority of usage has migrated to GPT‑5.x and that just 0.1% of daily users are still choosing GPT‑4o, and it framed the retirements as a way to focus on the models most people use today. [2][3] The decision follows earlier controversy when GPT‑4o was briefly removed after the GPT‑5 launch and then restored after strong user pushback; OpenAI leadership had previously said they would give “plenty of notice” before any deprecation, and the company also points to improvements in GPT‑5.1/5.2 (including personality and customization controls) as reasons users can migrate to newer models. Many users and community groups responded emotionally online — describing heartbreak and anger at losing a preferred conversational style — and criticized the timing and short notice window. OpenAI emphasized no immediate API changes and highlighted new personality controls as mitigation. [2][1][3] In conclusion, the announcement formalizes a move toward newer GPT‑5 generation models and customization features while exposing tensions between product lifecycle decisions and communities that form attachments to particular model personalities; expect continued user pushback focused on notice periods and preservation of specific behaviors, even as OpenAI encourages migration to GPT‑5.1/5.2 and related replacements. [2][3]
Controversy
Some users and community posts argue OpenAI reneged on an earlier promise to give “plenty of notice” before removing GPT‑4o and other models, criticizing the two‑week window as insufficient; OpenAI says usage has shifted to GPT‑5 and cites low daily use (0.1%) as justification for the retirement [2][3].