OpenAI CEO Sam Altman has announced the addition of ‘Auto’, ‘Fast’, and ‘Thinking’ modes to GPT-5, allowing

 ChatGPT users to choose how the AI prioritises speed and reasoning. The feature, however, is limited to paid subscribers. This update follows criticism that GPT-5’s performance had dipped compared to older models, prompting some users to call for the return of GPT-4o. GPT-5 originally used a real-time router to automatically select the optimal model for a task, but Altman admitted that this system sometimes reduced answer quality. The new mode selector gives users greater transparency and control: Auto balances speed and reasoning, Fast delivers quicker replies, and Thinking focuses on deeper analysis even if it takes longer. This change is part of OpenAI’s broader push to improve the ChatGPT experience.