When OpenAI launched GPT-5, it promised an all-in-one AI model capable of handling every type of task without users having to switch between models. The company even retired older options, including the popular GPT-4o. But instead of applause, the update has sparked frustration among ChatGPT’s loyal users.
Many have taken to social media and forums to say that GPT-5 feels less capable in certain areas. According to complaints, the model’s responses are shorter, less detailed, and missing the personality that made GPT-4o a favourite for creative and conversational work. Some users say they built specific workflows with GPT-4o that the new version simply cannot replicate.
During a Reddit AMA, one user voiced concern that GPT-5 might be better for coding and technical reasoning, but it falls short for creative collaboration and high-context roleplay. OpenAI CEO Sam Altman acknowledged the criticism and promised changes. Shortly after, he announced that Plus subscribers would soon get double the usage limits for GPT-5 and, importantly, continued access to GPT-4o for the time being.
Altman also revealed that a key feature called the “autoswitcher” which routes queries to the most suitable capability, whether for reasoning, creativity, or conversation had been inactive since launch. This, he believes, is why GPT-5 appeared less impressive. OpenAI is now fixing the feature and will also make the active model visible to users to remove any confusion.
Despite the backlash, Altman reiterated that the company is running at a loss, with billions spent annually on server costs. While older models like GPT-4o may remain for now, OpenAI faces a tough decision between controlling costs and keeping users happy. For now, the company is betting on GPT-5’s upcoming improvements to win people over.
For more updates on the latest AI developments and tech news, follow Tech Moves on Instagram and Facebook.
