GPT-Live 1 now available on AI Gateway
Full-duplex voice that doesn't wait for pauses. GPT-Live 1 is live on AI Gateway with client delegation.
Why it matters
GPT-Live 1 removes turn-taking latency in voice conversations and lets developers delegate to other models mid-conversation without interrupting the user. This is a meaningful UX shift for voice AI apps and changes how practitioners architect voice-first experiences.
The key facts
6 to knowGPT-Live 1 is full-duplex: listens and speaks simultaneously, eliminating turn-detection pauses
Client delegation allows independent background model selection while voice session continues
Available on AI Gateway via @ai-sdk/openai 4.0.67 or later with WebSocket client
Delegated model requests billed separately; voice-session usage continues during delegation
Users can pause, interrupt, or add detail while GPT-Live is speaking
Published Sep 17, 2026 (Vercel changelog)
Go to the source
Vercel Blogvercel.com
Publisher excerpt: is now available on AI Gateway. GPT-Live 1 from OpenAI GPT-Live 1 is a full-duplex voice model and can listen and speak at the same time. Many voice models use turn detection to respond. Full duplex removes that boundary, so a user can pause, interrupt, or add detail while GPT-Live is speaking.…