ChatGPT users can now adjust GPT-5 Thinking speed: How the new update works

OpenAI has unveiled an exciting update that allows ChatGPT users to adjust the response speed of GPT-5 with Thinking. This feature, revealed on September 18 via X, is available to Plus, Pro, and Business subscribers, enabling them to control the model's 'thinking time' directly from the message composer on the web. Previously, GPT-5 with Thinking was designed to take its time in generating responses, focusing on careful reasoning. However, user feedback indicated a desire for more flexibility to balance response depth and speed based on specific tasks. With this new update, subscribers can now choose from several pacing options: - **Standard**: The default mode for all plans, providing a balanced approach between speed and intelligence. - **Extended**: Previously the standard for Plus users, it allows the model to take more time for deeper, more thoughtful responses. - **Light**: Exclusive to Pro users, this mode delivers quick responses for less complex tasks. - **Heavy**: Also available to Pro users, this mode prioritizes in-depth reasoning, albeit at a slower pace. OpenAI has stated that the selected mode will remain active across future chats on the web until users decide to change it. The company acknowledged user concerns regarding response times, stating, "We’ve heard your feedback that GPT-5 Thinking can sometimes take longer than you’d like. Now Plus, Pro, and Business users can set the pace to match the moment." In addition to this update, OpenAI recently launched GPT-5 Codex, a new AI model tailored for software development purposes. This advanced version is designed to assist developers in writing, debugging, and reviewing code more efficiently, functioning both collaboratively with humans and autonomously on complex tasks. According to OpenAI, GPT-5 Codex can tackle a broad spectrum of programming challenges, from initiating projects to enhancing features, fixing bugs, and reviewing code created by others. The AI is capable of independently managing long-term projects, testing and refining its own code over several hours to ensure functionality. For simpler tasks, it provides quick responses without unnecessary resource consumption. OpenAI emphasized that the AI has been trained to meticulously review code, identifying critical errors before they lead to issues, enhancing the overall development process. Furthermore, tests on popular open-source projects have demonstrated that GPT-5 Codex offers more accurate and beneficial suggestions than ever, allowing developers to focus on the most pressing concerns.