OpenAI is officially phasing out several older AI models from the ChatGPT interface starting February 13, 2026, marking another major step in the platform’s rapid evolution. The models being removed include GPT 4o, GPT 4.1, GPT 4.1 mini, and OpenAI o4 mini. This follows the earlier retirement of GPT 5 Instant and GPT 5 Thinking variants. While these changes affect what users can select inside ChatGPT, API access to the models remains available for now.
The move signals a clear shift in focus as OpenAI concentrates its efforts on newer systems like GPT 5.2, which now powers most interactions across the platform.
GPT 4o, in particular, developed a loyal following. Known for its expressive tone, multilingual fluency, and multimodal capabilities, it became a favorite for users who valued its warmth and personality. After the rollout of GPT 5 last year, user backlash even prompted OpenAI to briefly bring GPT 4o back into the interface. Coverage from Digital Trends highlighted how strong that reaction was, especially from users who felt the newer models lacked the emotional nuance they had grown used to.
But usage patterns tell a different story. According to OpenAI’s official announcement on its website, only a small percentage of daily ChatGPT users were still actively choosing these older systems. As the majority transitioned to GPT 5.1 and GPT 5.2, maintaining legacy infrastructure became less practical.
From today onward, those older models will no longer appear in the ChatGPT model selector. Existing conversations that originally ran on a retired model will automatically shift to a newer one, such as GPT 5.2. For most users, the transition should feel seamless. Saved chats, custom GPTs, and workflows will continue functioning as usual, even though the underlying model powering them has changed behind the scenes.
There is, however, an emotional layer to this change. Social media platforms like X have seen users share posts expressing disappointment, particularly around GPT 4o’s perceived empathy and conversational style. Some described the model as supportive, patient, and uniquely reassuring during difficult moments. The hashtag keep4o gained traction among fans who hoped OpenAI might reconsider.
OpenAI has acknowledged that feedback from GPT 4o users influenced improvements in its newer releases. Personality controls and customization features introduced in GPT 5.1 and GPT 5.2 were shaped in part by the community response. The company has emphasized that while older models are being retired, their strengths have not been ignored. Instead, those qualities are being integrated into more advanced systems.
Beyond user sentiment, there are practical reasons for consolidating model offerings. Supporting multiple legacy AI systems requires ongoing maintenance, security updates, infrastructure resources, and safety monitoring. By narrowing the lineup, OpenAI can allocate more engineering time toward refining performance, improving response accuracy, enhancing safety safeguards, and advancing multimodal intelligence.
The broader context also highlights how quickly generative AI continues to move. In just a few years, models have progressed from text only chatbots to systems capable of processing images, audio, and complex reasoning tasks. As OpenAI competes with rivals such as Google’s Gemini and Anthropic’s Claude, the pressure to innovate remains intense. Concentrating development around fewer, more capable models allows for faster iteration and deeper improvements.
For everyday users, the biggest difference will simply be the absence of familiar names in the dropdown menu. Conversations will continue. Custom instructions will still apply. Productivity workflows, creative writing sessions, coding help, and research assistance will function much the same way they did before. The only noticeable shift may be subtle changes in tone or phrasing as GPT 5.2 handles requests that once ran on GPT 4o or GPT 4.1.
This update also reflects a broader pattern in technology platforms. Products evolve. Features change. Popular tools eventually give way to newer versions that promise greater speed, stronger reasoning, and improved reliability. In the world of AI language models, that cycle tends to move even faster.
For some longtime users, saying goodbye to GPT 4o feels personal. For OpenAI, it represents a strategic decision aimed at streamlining development and focusing on what comes next.
