With a blazing-fast ~200 output tokens per second, Nova Micro is the fastest model — perfect for high-traffic applications. It’s also the most cost effective model with half the price of the next cheapest model.

Source: Independent 3rd party via Artificial Analysis.