OpenAI has opened its GPT-Live-1.1 voice model API to third-party developers. Built on a full-duplex architecture, the model listens and speaks simultaneously while detecting pauses, interruptions, and pace changes. Its key advance is delegating deep reasoning and tool-calling tasks to a developer-chosen backend model without interrupting the live conversation, merging real-time voice interaction with multi-step agent workflows.
Full-duplex Voice Interaction
全双工语音交互
Real-time speech architecture that listens and speaks simultaneously, handling interruptions and pauses.
Curated Intelligence
Startup Orilo has launched a pure voice Personal Agent designed to address latency and interruption awkwardness in current AI assistants. The product emphasizes a 'human-like' natural conversation experience, allowing users to interrupt mid-sentence and assign new tasks without causing silence. While the article contains significant marketing rhetoric, the demonstrated real-time voice interaction capabilities represent a new direction for consumer-grade agents, echoing Steve Jobs' original vision for Siri.
PolyAI CTO Shawn Wen stated at HumanX that voice AI has not yet reached its 'ChatGPT moment' despite the advent of full-duplex models. He identified fast reasoning for natural conversation as the next hurdle and emphasized the need for transparency and trust in enterprise deployments.
ZTE Nubia announced that the second-generation Doubao phone, NaviX Ultra, will launch its 'Beta Lab' livestream on October 14. The device features an industry-first dedicated AI button with integrated fingerprint recognition and the latest Seed full-duplex voice model, enabling complex cross-app task execution via single voice commands.