OpenAI Brings Back GPT-4o After GPT-5 Backlash

OpenAI has reinstated access to its GPT-4o model for ChatGPT Plus subscribers after facing significant criticism over the recent rollout of GPT-5. The change follows widespread complaints about the new model’s conversational style and perceived decline in writing quality.

Earlier this week, OpenAI launched GPT-5, describing it as its most advanced system. The update removed the option for users to choose between models, automatically defaulting to GPT-5. The new release incorporates a “real-time router” that shifts between a more efficient model for simple requests and a deeper reasoning version for more complex tasks.

Shortly after the launch, online forums such as Reddit began filling with criticism from users who preferred GPT-4o’s responses. Many described GPT-5’s replies as “sterile,” “brief,” and lacking personality. Some users expressed an emotional connection to GPT-4o, saying they considered it a consistent and engaging assistant.

Altman Acknowledges Feedback

OpenAI CEO Sam Altman addressed the concerns during a Reddit AMA, stating that while GPT-5 performs better in most benchmarks, the company “underestimated how much some of the things that people like in GPT-4o matter to them.” Altman later confirmed on X that GPT-4o would be restored as an option for Plus subscribers.

Future of GPT-4o Remains Uncertain

Altman has not committed to keeping GPT-4o available indefinitely, saying the company will monitor usage before making a decision. In the meantime, OpenAI is focusing on completing GPT-5’s rollout and adjusting its tone to create what Altman described as a “warmer” conversational experience.

While the restoration of GPT-4o addresses immediate user concerns, the long-term direction of OpenAI’s model offerings remains unclear. For users who rely on GPT-4o’s style and consistency, the current reprieve may only be temporary.