OpenAI has started rolling out a new toggle feature in ChatGPT, allowing users to adjust the reasoning intensity of the GPT-5 Thinking model and providing greater control over its performance.
The company has been testing this functionality for several weeks, initially considering a slider interface but ultimately opting for a simpler toggle design. The toggle is now available to Plus and Pro subscribers, with Plus and Business users gaining access to the core options.
The toggle allows selection of different “juice” values, which represent internal reasoning effort levels. The default Standard mode operates at a juice value of 18, striking a balance between processing speed and intelligence. Users can also select Extended mode, set at juice 64, for enhanced reasoning capabilities.
Pro subscribers receive additional flexibility with Light mode at juice 5, designed for near-instantaneous responses, and Heavy mode at juice 200, which maximizes the model’s thinking depth. This rollout enhances user experience by tailoring the AI’s effort to specific needs, whether prioritizing quick replies or in-depth analysis.
According to reports, the feature aims to optimize the GPT-5 Thinking model’s utility across various tasks. OpenAI continues to refine its offerings, building on the model’s advanced capabilities introduced earlier this year.
