The Decreasing Value of Chain of Thought in Prompting

What it is

A June 2025 Wharton Generative AI Labs technical report by Lennart Meincke, Ethan Mollick, Lilach Mollick and Dan Shapiro. It tested "think step by step" prompting and found modest average gains with more variable answers for non-reasoning models, and only marginal gains for reasoning models despite 20–80% longer response times.

Topics