OpenAI engineered GPT-Live, a full-duplex realtime voice architecture that eliminates turn detectors and reduces transport latency.

Key Details

• Eliminates turn detector with a full-duplex model that listens and speaks simultaneously
• Delegates complex reasoning and tool execution asynchronously to GPT-5.5
• Rewrote media frontend in Go, bringing p95 latency down to previous p50 levels
• Developed WARP protocol with IETF to cut WebRTC startup handshake from six round trips to one

In This Story