OpenAI now lets you control how fast ChatGPT thinks
What's the story
OpenAI has rolled out a new feature for its ChatGPT service, giving users more control over the response time of GPT-5 with Thinking. The update, announced on X (formerly Twitter), is available for Plus, Pro and Business subscribers. It allows them to adjust the "thinking time" of the model directly from the message composer.
User control
Users can choose from 4 different pacing modes
The new update comes after many users complained about the slow response time of GPT-5 with Thinking. The model was designed to reason carefully before answering, often taking longer than expected. Now, subscribers can choose from four different pacing modes: Standard (default for all plans), Extended (previously default for Plus users), Light (exclusive to Pro users) and Heavy (also exclusive to Pro).
Persistent settings
Selected pacing mode will remain active across future chats
OpenAI has confirmed that the selected pacing mode will remain active across future chats on the web until users decide to change it manually. The company said, "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." This gives users more control over their interaction with ChatGPT.
Developer tool
OpenAI also introduces GPT-5 Codex for coding tasks
Along with the new pacing feature for ChatGPT, OpenAI has also launched GPT-5 Codex. This AI model is specifically designed for software development and can help developers write, debug, and review code more efficiently. It can handle a wide range of programming tasks from creating projects from scratch to adding features or fixing bugs in existing ones.