OpenAI Voice Model Upgraded to Full-Duplex

APPSO · wechat · 2026-07-09

This post introduces OpenAI's new voice model GPT-Live, highlighting its upgrade from turn-based voice interaction to a full-duplex architecture that can listen and speak simultaneously. It reportedly provides natural interjections like 'uh-huh' or 'got it', waits when users hesitate, and improves recognition and real-time translation in noisy environments.

Related event: OpenAI Launches Full-Duplex Voice Model GPT-Live(44 posts)→

Original post →

More from Models

Models channel →