OpenAI Introduces Flexible Thinking Controls in ChatGPT

OpenAI has unveiled a major update to ChatGPT, giving users the ability to choose how much reasoning effort the GPT-5 Thinking model invests before producing a response. This new control feature is designed to balance speed with depth, allowing users to tailor the AI’s output depending on their task.

The update is being gradually rolled out to Plus, Pro, and Business subscribers, offering different access levels depending on the plan.

Four New GPT-5 Reasoning Modes

The new feature introduces four distinct modes, each representing a different level of “thinking effort”:

  • Light Mode: Very fast, minimal reasoning (great for quick answers).

  • Standard Mode: Balanced response speed and reasoning depth; default for most users.

  • Extended Mode: More detailed and thoughtful answers, with longer wait times.

  • Heavy Mode: Maximum reasoning depth, ideal for complex problems but slowest to deliver results.

While OpenAI internally refers to “juice” values (a measure of computational effort), end users only see these mode names.

Availability Across ChatGPT Plans

  • Plus & Business users: Access to Standard and Extended modes.

  • Pro users: Full access, including Light and Heavy modes.

Once selected, a mode stays active across chats on the web until changed manually, making it easy for users to stick with their preferred setting.

Why This Matters

This move reflects OpenAI’s push toward greater personalization and efficiency in AI usage. Different tasks require different AI behaviors:

  • Speed-first tasks (like drafting emails or generating short notes) work best with Light or Standard modes.

  • Deep reasoning tasks (like analyzing data, writing code, or solving complex technical problems) benefit from Extended or Heavy modes.

By letting users decide, OpenAI is addressing complaints about slow response times while maintaining the option for powerful, detailed reasoning when needed.

Industry Reaction and Future Outlook

Experts see the update as part of a broader trend: AI models becoming more transparent and user-centric. Rather than offering a one-size-fits-all solution, ChatGPT now adapts to the user’s pace and requirements.

However, heavier reasoning consumes more computational resources, raising questions about scaling and long-term costs. OpenAI’s decision to reserve Light and Heavy modes for Pro users may be a strategic way to balance accessibility with infrastructure demands.

The Bottom Line

OpenAI’s new thinking depth controls put users firmly in charge of ChatGPT’s performance. Whether you need quick replies or in-depth analysis, GPT-5 can now adapt to your workflow—making the tool more versatile than ever before.