Tag
From the Library
Library
Inference-time compute (Reasoning models)
Extra computation for an individual prompt can improve some demanding tasks, but it does not guarantee a correct answer and adds latency.
