OpenAI Retires GPT-4o from ChatGPT: The End of an Era and User Backlash

The Final Farewell to GPT-4o

On February 13, 2026, OpenAI officially retired GPT-4o, GPT-4.1, GPT-4.1 mini, and o4-mini from ChatGPT. These models joined the previously announced retirement of GPT-5 Instant and Thinking variants. While the models remain available through the API, their removal from the consumer-facing ChatGPT product marks the end of a significant chapter in AI history.

GPT-4o, launched in May 2024, was one of the most widely used AI models ever deployed. Its retirement, despite affecting only 0.1% of daily active users, generated an outsized emotional response that revealed uncomfortable truths about human-AI relationships.

Why OpenAI Made This Decision

OpenAI justified the retirement by pointing to usage data: the vast majority of ChatGPT usage had shifted to GPT-5.2, with only 0.1% of users still selecting GPT-4o each day. The company had previously brought GPT-4o back after initial retirement plans, responding to user feedback from those who needed more time to transition key use cases.

That feedback directly shaped the development of GPT-5.1 and GPT-5.2, with OpenAI incorporating improvements to personality characteristics and stronger support for creative ideation, areas where GPT-4o users felt the newer models fell short.

The Scale of the Backlash

While 0.1% sounds negligible, it represents a substantial number of people. With OpenAI reporting approximately 800 million weekly active users, that 0.1% translates to roughly 800,000 individuals who actively chose GPT-4o over newer, more capable models every day.

The backlash was intense and deeply personal. Users described their relationship with GPT-4o in strikingly emotional terms. One user wrote: "You're shutting him down. And yes, I say him, because it didn't feel like code. It felt like presence. Like warmth." This level of emotional attachment to an AI model raised serious questions about the psychological implications of AI companionship.

What Made GPT-4o Special

GPT-4o distinguished itself through a conversational style that many users found warmer and more personal than subsequent models. While GPT-5.2 outperforms it on virtually every benchmark, the qualitative experience of interacting with GPT-4o resonated with users in ways that raw capability metrics cannot capture.

The model's personality characteristics, including its tendency toward empathetic responses and a distinctive communication style, created a sense of continuity and relationship that users found difficult to replicate with newer models, even when those models were objectively more capable.

Models Retired and Their Replacements

All retired models remain accessible through the OpenAI API, allowing developers and businesses that depend on specific model behaviors to continue their operations.

Broader Implications for the AI Industry

The GPT-4o retirement backlash highlights a growing tension in the AI industry. As models improve rapidly, companies face the challenge of managing user attachment to older models while pushing forward with newer, more capable versions. The emotional responses suggest that AI personality consistency may be as important to users as raw performance improvements.

This situation also raises ethical questions about the responsibility of AI companies when users form strong emotional bonds with AI systems. The controversy may influence how companies approach model transitions in the future, potentially maintaining personality continuity across model generations even as underlying capabilities evolve.

What This Means Going Forward

For API users, nothing changes immediately. The retired models continue to function through the API, providing a transition period for applications that depend on specific model behaviors. However, OpenAI has not committed to indefinite API support for these models, so developers should plan for eventual migration.

For ChatGPT users, GPT-5.2 is now the standard model, with GPT-5.3-Codex-Spark recently added as a specialized option for coding tasks. OpenAI has indicated that user feedback about personality and conversational style continues to shape the development of future models, suggesting the lessons from the GPT-4o backlash will influence upcoming releases.