How we built a realtime system for responsive voice AI in six months
GPT‑Live is a third-generation voice system that eliminates the need for a turn detector, enabling full-duplex conversations with low latency. It uses a stateful inference architecture to maintain context and handle continuous inference, ensuring sub-second responsiveness. The system supports asynchronous delegation, dynamic context compaction, and efficient media routing, allowing it to consult frontier models like GPT‑5.5 without interrupting the conversation.
More from Openai
How RingCentral builds AI-native work from engineering to ops
See how RingCentral uses ChatGPT Work and Codex to accelerate AI product development and centralize operational intelligence across engineering and operations.
Daybreak models are now available on AWS
OpenAI and AWS are making Daybreak cybersecurity capabilities available through Amazon Bedrock to support enterprise security workflows.
Testing ads in ChatGPT
OpenAI begins testing ads in ChatGPT to support free access, with clear labeling, answer independence, strong privacy protections, and user control.
