Skip to content

Topic

Reasoning token efficiency

How many chain-of-thought tokens a model spends to reach a correct answer, and the techniques used to shorten that trace without losing accuracy.

Current clusters