^{Photo: CNN}

OpenAI to Remove Several Legacy Models From ChatGPT

OpenAI announced this week that it will retire multiple models from its ChatGPT platform next month, including GPT-4o — a conversationally warm and widely appreciated version that debuted in May 2024.

The move is part of a broader platform overhaul as usage data shows only a tiny fraction of users continue to select GPT-4o on a daily basis. OpenAI says this transition will help focus development on the models most widely used and best able to support advanced AI tasks.

In its announcement, the company acknowledged that some users will be disappointed, particularly fans of GPT-4o, but emphasized that retiring older models is necessary to maintain system efficiency across millions of global users.

Why OpenAI Is Retiring Models

GPT-4o was introduced with fanfare for its natural dialogue style and creative responses, quickly gaining popularity among a segment of paying subscribers. After the launch of GPT-5 last year, OpenAI briefly removed access to GPT-4o, sparking user feedback and a subsequent reversal that restored access for paid accounts.

At the time, CEO Sam Altman pledged to give “plenty of notice” if the model were ever to be fully retired. That notice has now arrived.

OpenAI explained that daily usage metrics show only roughly 0.1% of users choose GPT-4o, with the vast majority gravitating toward GPT-5.2 — the more powerful successor optimized for general usage across tasks including reasoning, code generation, and creative writing.

In recent software updates, OpenAI has also introduced enhancements to model personality features, customization tools, and ideation capacities, reducing the need for older variants.

Models on the Chopping Block

In addition to GPT-4o, OpenAI confirmed that the following models will be phased out from ChatGPT next month:

• GPT-4.1
• GPT-4.1 Mini
• OpenAI o4-mini
• Previously announced removals of GPT-5 Instant and GPT-5 Thinking will also proceed

These retirements are focused exclusively on the ChatGPT user interface and do not affect OpenAI’s application programming interface (API). Developers integrating models programmatically through the API will continue to have access to supported engines without changes at this time.

What This Means for Users

For people who enjoy the nostalgic tone or specific response style of GPT-4o, the change may feel significant. OpenAI stressed that model removal is never taken lightly and that longstanding features like personalization and conversation memory will be preserved and improved within the remaining model lineup.

Users are encouraged to transition their workflows and preferences to GPT-5.2, which OpenAI says delivers better performance in most conversational and creative scenarios.

While the company did not specify plans for future model releases beyond the current lineup, analysts believe this model consolidation could signal a more unified AI experience where performance, customization, and reliability are the top priorities.

Looking Ahead: Focus on Future AI Development

By trimming legacy models, OpenAI aims to streamline support, reduce infrastructure complexity, and free up resources to accelerate research and development in its most advanced engines.

This approach mirrors broader industry trends in AI, where companies iterate quickly on core models while retiring older versions that see diminishing usage.

For users, the transition may also deliver benefits such as faster responses, more consistent insights, and enhanced features powered by the latest AI architecture.

As the phase-out date approaches, OpenAI will likely provide more guidance on migration paths and tips for users moving from retired models to newer ones, ensuring continuity of experience within ChatGPT’s expanding ecosystem.

‍