OpenAI brings back GPT-4o after user revolt (arstechnica.com)

OpenAI has reinstated GPT-4o to ChatGPT following a strong user backlash after its removal during the recent GPT-5 rollout. OpenAI CEO Sam Altman acknowledged underestimating how much users valued GPT-4o’s features, which sparked one of the most vocal user revolts in ChatGPT’s history, with thousands expressing frustration over losing access to familiar and preferred models. The GPT-4o model now appears by default in the model picker for all paid ChatGPT users, including ChatGPT Plus subscribers, signaling a quick pivot in response to community feedback. Alongside GPT-4o’s return, OpenAI introduced enhancements aimed at improving user experience and flexibility. Rate limits for the premium GPT-5 Thinking mode increased significantly from 200 to 3,000 messages per week, with additional usage available via the new GPT-5 Thinking mini. Furthermore, users can now select routing options like “Auto,” “Fast,” and “Thinking” to better customize how GPT-5 variants process their queries. For Pro users paying $200 monthly, OpenAI plans to offer access to more legacy and specialized models—including o3, 4.1, and GPT-5 Thinking mini—through an upcoming toggle in settings, while GPT-4.5 remains a Pro-exclusive due to its intensive GPU requirements. This move highlights OpenAI’s responsiveness to user preferences and its effort to balance model innovation with community needs.