OpenAI has officially announced the retirement of multiple older AI models from ChatGPT, with GPT-4o emerging as the most widely discussed name on the list. The model is scheduled to be removed from ChatGPT on February 13, bringing an end to a long and often debated presence on the platform. The decision follows an earlier attempt to phase out GPT-4o that was reversed after intense user backlash.
Many users had argued that newer models failed to match GPT-4o’s emotional nuance and writing quality. This time, however, OpenAI says the situation has changed.
Why OpenAI Is Retiring GPT-4o from ChatGPT
OpenAI confirmed that GPT-4o will be removed from ChatGPT alongside GPT-5 Instant, GPT-5 Thinking, GPT-4.1, GPT-4.1 mini, and o4-mini. The company clarified that the move applies only to ChatGPT users, while developers accessing OpenAI’s API will continue to use these models without interruption.
GPT-4o earned a loyal following for its natural tone, emotional intelligence, and expressive writing style. When OpenAI initially attempted to retire the model in August 2025 following the launch of GPT-5, users strongly pushed back, claiming the newer model felt overly mechanical and lacked warmth. Responding to that feedback, OpenAI reinstated GPT-4o.
ALSO READ: Sony PS Plus January 2026 Games List: Need for Speed, Disney Epic Mickey, Core Keeper
According to OpenAI, usage patterns have since shifted. The company says that only 0.1 percent of daily ChatGPT users now actively select GPT-4o, leading it to conclude that maintaining the model is no longer justified.
GPT-5.2 Update and Renewed Criticism
OpenAI maintains that the qualities users valued in GPT-4o have now been integrated into GPT-5.2. These improvements reportedly include enhanced personality tuning, greater creativity, and expanded customization options that allow users to adjust response tone, warmth, and enthusiasm.
Despite these claims, the announcement has reignited criticism online. Several users on X questioned OpenAI’s usage statistics, arguing that GPT-4o engagement declined because GPT-5.2 became the default model while GPT-4o was placed behind a paywall. Others pointed out that opposition to the decision quickly gained traction, with thousands of posts expressing disapproval.
The debate intensified further after OpenAI CEO Sam Altman acknowledged that GPT-5.2’s writing quality had suffered as the company prioritized improvements in coding, reasoning, and mathematical performance. For many users, this admission has made GPT-4o’s retirement feel less like progress and more like a step backward.