OpenAI Responds to User Demand for GPT-4o

In a surprising turn of events, OpenAI has decided to bring the much-loved GPT-4o back online after an uproar among users following the launch of its new model, GPT-5. Many devoted fans of GPT-4o took to social media to express their deep dissatisfaction with the upgraded model, insisting that they missed what they fondly referred to as their 'AI friend'.

What Went Wrong with GPT-5?

The rollout of GPT-5, branded by OpenAI as the "smartest, fastest, and most useful model yet," was met with backlash. Unlike its predecessor, the new model defaults to GPT-5 and removes the option for users to select their preferred version. Many users reported that GPT-5 seemed less capable, with OpenAI's CEO Sam Altman admitting it 'felt way dumber' in certain respects.

A Troubling Launch and User Criticism

In a candid Reddit AMA, Altman addressed concerns over the new model's performance, acknowledging that while GPT-5 is designed to perform better overall, it may have missed the mark on writing quality. Users described GPT-5's output as "sterile" and "much worse," preferring the more engaging style that characterized GPT-4o.

OpenAI's Decision: A Return to GPT-4o

Following the overwhelming feedback, Altman took to X (formerly Twitter) to announce a significant change: OpenAI will allow Plus subscribers to choose between using GPT-5 or returning to the beloved GPT-4o. He admitted that the company underestimated the importance of certain features that users appreciated in GPT-4o, even noting that GPT-5 performs better in many regards.

The Community Speaks: What Does This Mean?

This move highlights the power of user feedback in shaping technology. As OpenAI navigates this landscape of AI development, the revival of GPT-4o serves as a reminder that sometimes, the heart holds more weight than mere performance metrics. It will be interesting to see how both models evolve and how users react as GPT-5 continues to be refined.