OpenAI’s much-hyped GPT-5 mannequin launched on Thursday, however not like what the ChatGPT maker would have anticipated, its newest and best mannequin was met with a whole lot of scrutiny on social media, with longtime customers of the chatbot sad with the corporate over the elimination of older OpenAI fashions like GPT-4o and GPT-4.1. Notably, after the launch of GPT-5, OpenAI stopped supporting the older fashions and GPT-5 grew to become the default mannequin for all duties, resulting in widespread frustration on social media.
OpenAI has claimed marked enhancements within the efficiency of ChatGPT with GPT-5, together with reasoning, coding, accuracy, multimodal understanding, writing, and answering health-related questions. Nonetheless, customers on Reddit more and more complained that the brand new mannequin has much less persona than GPT-4o and tends to offer shorter solutions to queries.
Many customers identified how GPT-4o was used as a companion, and whereas GPT-5 might have an edge in technical duties, it doesn’t have the identical sort of responses because the older, in style mannequin. As one Reddit person identified, “I do know GPT-5 is designed to be stronger for complicated reasoning, coding, {and professional} duties, however not all of us want a professional coding mannequin. A few of us depend on 4o for artistic collaboration, emotional nuance, roleplay, and different long-form, high-context interactions. These areas really feel completely different sufficient in GPT-5 that it impacts my skill to work and create the best way I’m used to.”
An OpenAI researcher, Kristina Kim, additionally admitted throughout a Reddit AMA session that OpenAI has “made a devoted effort with GPT-5 to coach our mannequin to be extra impartial by default.”
Sam Altman confirms OpenAI is bringing again GPT-4o
In the meantime, OpenAI CEO Sam Altman gave in to the demand by ChatGPT customers to carry again GPT-4o, however just for the corporate’s Plus subscribers. Altman didn’t fully backtrack on the retirement announcement made by OpenAI however said that they might watch utilization of GPT-4o to find out how lengthy to assist the mannequin.
In the course of the Reddit AMA, Altman wrote, “okay, we hear you all on 4o; thanks for the time to offer us the suggestions (and the eagerness!). we’re going to carry it again for plus customers, and can watch utilization to find out how lengthy to assist it.”
Notably, ChatGPT Professional, Workforce, and Enterprise customers already at present have an choice to get again to older fashions by turning on the “present legacy fashions” banner within the settings. This, paired with Altman’s announcement, means it is kind of the top of the highway without spending a dime customers so far as GPT-4o and different older fashions are involved.
In response to a different person’s query on Reddit, Altman requested if it is vital for them to regain entry to GPT-4.1 or if simply GPT-4o would suffice, giving a potential indication that OpenAI can also carry again GPT-4.1 if there’s sufficient demand for the mannequin amongst its Plus customers.
