OpenAI Didn’t Publish GPT-Live’s Latency. So We Measured It.

OpenAI's GPT-Live, launched on July 8, 2026, aims to enhance AI-powered voice interactions by introducing a full-duplex mode that listens and responds simultaneously, unlike its predecessors which operated on a turn-based system. Agora Media Lab conducted tests revealing GPT-Live's ability to handle interruptions and background speech more effectively than earlier models, although it occasionally misattributed background voices as user input in noisy environments. The new model showed improved consistency in response times, reducing jitter significantly, which contributes to a perception of naturalness in conversations. GPT-Live also demonstrated resilience to network issues, outperforming previous models even under 10% packet loss. However, achieving a seamless user experience requires synchronization across the model, client, network, and playback, which OpenAI manages by controlling the entire stack, unlike most teams that rely on a model API over the public internet. The findings highlight the importance of consistent timing and network conditions in delivering natural full-duplex voice interactions.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.