Have you noticed that we now display the reasoning iterations for applicable models?