Control Reasoning Effort in Playground and better feedback on thinking models

August 13, 2025

You can now fine-tune how models think and respond with the reasoning effort parameter.

What’s New

• Reasoning Effort Control
  • Minimal Reasoning: Added “minimal” option for faster, more direct responses
  • Low/Medium/High Reasoning: Existing reasoning levels for different complexity needs
• Visual Reasoning Display: See the model’s thinking process when available

Try the new reasoning controls in the Playground to see how different settings affect model responses.

Coming Soon

• Ability to set verbosity (low, medium, high) in the playground for GPT-5 models