OpenAI backs off retirement of 4o after backlash

OpenAI vows not to retire older models without warning after backlash over GPT-5 rollout, signaling a shift toward predictability for users.

ChatGPT keeps old models after GPT-5 backlash

OpenAI reintroduced 4o as an opt-in option for paying ChatGPT users after strong public backlash to the move to GPT-5. Nick Turley, OpenAI’s head of ChatGPT, told The Verge that retiring 4o first was a misstep and that many users formed a strong attachment to the model’s personality. CEO Sam Altman also weighed in, saying the update should feel not as annoying to most users as GPT-4o. The company says the change aimed to simplify model choices for its roughly 700 million weekly users, many of whom rely on the default option rather than selecting a model themselves.

Going forward, ChatGPT will not drop existing models without notice, Turley said, and OpenAI plans to extend the same predictability it uses for enterprise plans to more users. Despite the controversy, overall usage reportedly rose after GPT-5’s release, underscoring a divide between vocal power users and broad consumer audiences who are interacting with a new form of reasoning in AI tools.

The episode exposes a clash between product simplicity and emotional attachment to AI personalities. By prioritizing a single, default experience, OpenAI aimed to reduce cognitive load for casual users, but many users read 4o as part of the tool’s character. The company now signals a careful balance: keep innovations moving while preserving a sense of stability and awareness about upcoming changes. This move could become a template for other tech firms facing backlash when releasing new capabilities to millions of users.

For investors and enterprise customers, the shift matters because it hints at how quickly a company will introduce or retire features while maintaining predictable service. If predictability becomes a standard, firms may prioritize gradual rollouts and clearer notices over rapid, radical changes that disrupt daily workflows. The challenge ahead is maintaining trust as products grow more capable and the pace of change accelerates.

Highlights

Change will be watched closely by users and regulators alike.