Native interaction models: why 0.4s latency matters
Thinking Machines' native interaction models hit 0.40s latency, outpacing GPT-realtime2's 1.18s for concurrent audio and text streams.
Priya Nair
11 min read