OpenAI has rolled out user-controllable reasoning levels in ChatGPT, giving eligible subscribers a more direct way to choose how much reasoning a response should receive. The update is built around GPT-5.6 Sol and makes the trade-off between speed and more thorough work more visible in the ChatGPT interface.
Rather than leaving reasoning behavior entirely to automatic model selection, users can select a level from the model picker and manage automatic reasoning switching through a Configure option. The change matters because it turns an underlying model behavior into a practical, per-response decision: a user can favor a fast answer for a simple request or select a higher-reasoning option for work that warrants more effort.
How ChatGPT's reasoning controls work
According to OpenAI's ChatGPT help article on GPT-5.6 and GPT-5.5, the model picker now presents an Instant option plus Medium, High, Extra High, and Pro reasoning choices. Instant is powered by GPT-5.5 Instant, while Medium, High, and Extra High are powered by GPT-5.6 Sol. The Pro option uses GPT-5.6 Sol Pro for more difficult, longer-running tasks.
The interface also includes a Configure setting that lets users turn automatic reasoning switching on or off. That distinction is important. Automatic switching allows ChatGPT to manage the selection process, while manual selection gives the user a clearer instruction about the desired level of reasoning for a particular response.
