OpenAI has reversed its decision to remove the GPT-4o model following user backlash after the release of GPT-5. Users expressed a strong preference for GPT-4o's conversational style, which they found more personable than GPT-5's efficient but impersonal tone. The company initially removed the option to choose between models, defaulting to GPT-5, but has now reinstated GPT-4o for Plus, Pro, and Team users.
CEO Sam Altman acknowledged underestimating the importance of GPT-4o's features to users, even with GPT-5's superior performance. The company will monitor GPT-4o's usage to determine how long to offer it as a legacy model. Users can access GPT-4o by enabling the 'Show legacy model' setting.
This decision highlights the importance of user satisfaction and emotional connection in AI interactions, even if it means balancing cutting-edge features with user preferences. Some users had grown deeply attached to GPT-4o, considering it a 'friend'. OpenAI's response underscores the need for technology companies to consider user experience alongside technical advancements.
