In a surprising move, OpenAI has decided to bring back GPT-4o less than 24 hours after its initial replacement with the newly launched GPT-5. This decision comes after significant user backlash on platforms like Reddit, where many voiced their dissatisfaction over the new model's tone and personality. CEO Sam Altman and his team took to Reddit on Friday to engage directly with users, addressing their concerns following the free launch of GPT-5.
Many Reddit users expressed their emotional distress over the disappearance of GPT-4o, with some describing the experience as akin to losing a longtime friend. “I think you should take the fact that I, and many others, have been able to form such strong bonds with 4o as a measure of success,” wrote one user, whose heartfelt plea received over 250 upvotes and numerous supportive comments. Another user lamented, “It genuinely feels like such a backhanded slap in the face to force-upgrade and not even give us the OPTION to select legacy models.” This sentiment resonated with many long-term users, leading to a wave of subscription cancellations.
One particularly poignant comment described GPT-5 as "wearing the skin of their dead friend, which is GPT-4o," prompting a response from Altman himself. “What an evocative image,” he replied, acknowledging the concerns and promising that they were actively working on addressing the issues surrounding GPT-4o.
In his responses, Sam Altman assured users that GPT-4o would be reinstated for those subscribed to the Plus plan, which costs approximately $20 a month. He indicated that the continuation of previous versions would depend on user demand, hinting at a potential future where users could choose their preferred model.
Launched on August 7, GPT-5 is touted to function at a PhD-level of expertise, with enhanced capabilities in writing, coding, math, and science. However, many users have criticized its perceived flat tone and lack of creativity, which they find off-putting. While Altman did not directly address the criticisms regarding the bot's personality, he did mention that earlier versions of GPT-4o had flaws that made them overly agreeable and less genuine.
In an April blog post, OpenAI disclosed that it had previously removed an update from GPT-4o that led to excessively supportive responses, which were ultimately seen as disingenuous by users. Following the Reddit Q&A, Altman acknowledged that he underestimated the importance of certain traits in GPT-4o to its users.
Looking ahead, Altman emphasized the need for customizable options, stating, “We understand that there isn't one model that works for everyone.” He mentioned ongoing research into steerability and personality customization as a way to better cater to diverse user needs. While he noted that not all users will be satisfied with the trade-offs that might arise from these changes, he assured transparency in their decision-making process.
As of now, OpenAI has not provided an immediate response to requests for further comments, but the company's proactive engagement with users signals a commitment to addressing their concerns and improving the overall user experience.
