OpenAI Lets Users Control GPT-5 Thinking Speed for Faster or Deeper Responses

OpenAI has rolled out a new feature for ChatGPT that gives users the power to choose how long GPT-5 thinks before replying. This is part of the updates to the GPT-5 Thinking model.

Previously, when using GPT-5 “Thinking”, the model would often take longer by default in order to produce more detailed, thoughtful responses. Many users said sometimes it was too slow for simpler tasks. mint

Who Gets It and What Options Are Available

• The speed toggle is available to Plus, Pro, and Business subscribers.
• Depending on the plan, users have different mode options: PlanAvailable ModesPlus / BusinessStandard (default), Extended ProLight (fastest), Standard, Extended, Heavy (most in-depth)
• The “Standard” mode tries to balance speed and quality. “Extended” gives more time for deeper reasoning. “Light” is for snappier replies, and “Heavy” is for maximum depth.

How It Works in Practice

• When you select GPT-5 Thinking in the model selector (on web), a toggle appears in the message composer to choose your thinking speed mode.
• Once you choose a mode (e.g. Light, Heavy), that preference is remembered across future chats until you change it.
• For users who need quicker, more lightweight responses (e.g. fact lookup, quick brainstorming), the faster modes help. For tasks that need detailed reasoning (analysis, long explanations, complex problems), the slower/deeper modes are better.

Why This Matters

1. Better User Experience
   Users can now decide: do I want speed or depth right now? This helps with different tasks — quick answers vs complex ones.
2. Flexibility & Efficiency
   Not every query needs heavy thinking. Selecting faster modes can save time. Also, saving compute on simple tasks might help performance and cost.
3. Addressing User Feedback
   Many users had complained that GPT-5 Thinking was too slow for many interactions. This update is clearly a response to that.
4. Differentiation Among Subscription Tiers
   This adds another dimension to the Plus vs Pro vs Business offerings: not just how much you get, but how finely you can control the model’s behavior.

Things to Watch / Limitations

• The feature currently appears to be web-only. Mobile availability is mentioned as “coming soon”.
• Users still need a paid tier (Plus, Pro, Business) to access these toggles. Free users don’t seem to have the same control. TechRadar
• Even in “fast” modes, there may be trade-offs: depth, precision, nuance may suffer.

Conclusion

OpenAI’s new speed control feature for GPT-5 Thinking is a smart move toward giving users more control over their interaction with AI. It recognizes one size doesn’t fit all: sometimes speed matters more; other times accuracy and depth do. For paid users, this is now a useful lever to pull, depending on the task.