Sam Altman: OpenAI to Restore GPT-4o After User Outcry


Following a “rough” rollout of GPT-5, OpenAI CEO Sam Altman revealed the reinstatement of GPT-4o for paying ChatGPT users, with a stipulation. In a Reddit AMA and on X, Altman indicated that GPT-4o would once again be available for Plus users, permitting them to opt to continue its use while OpenAI assesses usage to determine the duration for which legacy models will be available.

OpenAI introduced GPT-5, vowing a more intelligent and quicker model with an optimized system for selecting the best version based on user prompts. Nonetheless, users criticized GPT-5 on Reddit and social platforms, labeling it a “failure” due to its less warm “personality” and the exclusion of other models from the ChatGPT selection. Despite earlier complaints about an excess of model choices, users sought the reintroduction of GPT-4o.

In the Reddit AMA, Altman reassured users that GPT-5 would rapidly improve, resolving problems with the autoswitcher and decision boundaries to better model selection. Users showed a preference for the warmth and empathy of GPT-4o, even with its flattering tendencies, which OpenAI had previously tackled. Some users expressed frustration over the short notice regarding the phase-out of prior models, although GPT-5 was designed to take their place.

Altman highlighted the aim of ensuring the new system would be more advantageous than earlier models and recognized the significance of effective change management.