Response time averages 350 milliseconds up to 700, it can sing, it can imitate voices, and its voice model is essentially indistinguishable from a human