OpenAI Restores GPT-4o After User Outcry Over GPT-5’s Quality

OpenAI recently launched its new GPT-5 model, but the rollout didn’t go as smoothly as planned. Many users immediately missed the previous GPT-4o, strongly demanding its return. Complaints poured in about GPT-5’s writing quality, with some users even expressing a sense of loss, describing GPT-4o as a trusted friend and confidant.

Touted as the “smartest, fastest, most useful model yet,” GPT-5 was introduced with a default setting, removing the option for users to choose an older model. OpenAI explained that GPT-5 employs a “real-time router,” designed to efficiently switch between a lighter model for simple queries and a more robust option for complex tasks.

However, this grand launch faced immediate backlash. Even OpenAI CEO Sam Altman admitted that the initial version of GPT-5 seemed “way dumber.” He engaged with users on Reddit, asking if they found GPT-5’s writing quality superior. The feedback was overwhelmingly negative: Redditors described GPT-5’s responses as “sterile,” “much worse,” and often “briefly and dryly.” Responding to the widespread user outcry, Altman quickly announced on X (formerly Twitter) that Plus subscribers would once again have the option to choose between GPT-5 and GPT-4o.

Altman acknowledged on X, “We for sure underestimated how much some of the things that people like in GPT-4o matter to them, even if GPT-5 performs better in most ways.”

The return of GPT-4o was met with widespread celebration, yet its long-term availability remains uncertain. Altman stated that OpenAI would “watch usage as we think about how long to offer legacy models for.” The company’s immediate focus is on completing the GPT-5 rollout and refining it to “make it warmer.” For the many users who have developed a strong connection with GPT-4o, viewing it as more than just a chatbot, this situation might signal the eventual farewell to their preferred AI companion.