OpenAI Retires GPT-4o and Legacy Models from ChatGPT Next Month

OpenAI Phases Out Older ChatGPT Models

OpenAI periodically updates its ChatGPT platform by retiring older models from the model selector as users shift to more advanced versions. On February 13, several legacy models, including the popular GPT-4o, will no longer be available in the interface.

Permanent Retirement of GPT-4o

Last August, OpenAI launched GPT-5 and promptly removed GPT-4o from the ChatGPT model options. This decision sparked significant user backlash, with many expressing preference for GPT-4o’s warmer, more engaging response style compared to GPT-5’s direct approach. The outcry highlighted ethical concerns and users’ emotional connections to large language models.

In response, OpenAI restored access to GPT-4o shortly after. OpenAI CEO Sam Altman addressed the issue, emphasizing the company’s commitment to user preferences. Since then, enhancements have included seven predefined personalities and options for custom instructions, allowing greater personalization of interactions.

Now, OpenAI plans to retire GPT-4o permanently, noting that it accounts for just 0.1% of daily user selections. This move aligns with efforts to streamline the platform toward newer, more capable models.

Additional Models Affected

Beyond GPT-4o, OpenAI will phase out several other older models from the ChatGPT selector. While the exact list includes legacy variants, the company confirms no immediate changes to the API, ensuring developers retain access for integrations.

These updates reflect OpenAI’s ongoing evolution of ChatGPT to prioritize efficiency and innovation while maintaining backward compatibility for essential services.