A new wave of voice intelligence

Three new Realtime API models, demonstrated live.