Tuesday 15 September 2026
Evening EditionRSS

ZOTPAPER

News without the noise

ZOTPAPER
AI

OpenAI launches full-duplex voice model GPT-Live-1 as API

Developers can now build simultaneous speech listen-and-respond features into apps, with improved interruption handling and lower latency

edit
By LineZotpaper
Published
Read Time2 min
OpenAI has released GPT-Live-1, its real-time full-duplex voice model, as an API, enabling developers to integrate simultaneous speech listening and generation into applications. First introduced in July through the ChatGPT interface, the model allows fluid conversation by handling interruptions naturally — a key improvement over earlier turn-based systems. Pricing starts at $0.05 per minute of voice usage, plus backend model costs.

The GPT-Live-1 API lets software developers build more responsive voice-enabled applications, from customer service bots to language tutors. Unlike half-duplex models that require users to take turns speaking and listening — akin to walkie-talkies — GPT-Live-1 can process input and output concurrently, enabling natural back-and-forth.

OpenAI highlighted early results from language-learning platform Speak, which found that GPT-Live-1 gave learners more time to think before the AI tutor responded, reducing interruptions by almost 80% compared to previous turn-based systems.

In benchmark testing, GPT-Live-1 scored 86.2% on the Tau3 voice-agent intelligence benchmark, which evaluates spoken customer-service tasks for airlines, retailers, and telecom firms, well ahead of GPT-Realtime-2.1 (45.7%) and GPT-Realtime-2 (42.4%). OpenAI said its own evaluations show GPT-Live-1 performing 30% better than GPT-Realtime-2.1 in Full Duplex Bench performance, with improved turn-taking latency and interactive behavior.

Yelp has already deployed GPT-Live-1 in its Yelp Host service for restaurant reservation management. "With GPT-Live-1, every call is more conversational and responsive," said Akhil Kuduvalli Ramesh, Yelp's chief product officer, in a statement. "Yelp Host now gives restaurants an advanced voice AI that delivers exceptional guest experiences, captures revenue opportunities they may have otherwise missed, and helps staff focus on serving guests instead of answering the phone."

The model is available through OpenAI's enterprise agent platform, OpenAI Presence. Custom voice options can be arranged through the sales team.

§

Analysis

Why This Matters

  • Business impact: GPT-Live-1 could transform customer service, restaurant booking, education and other industries where natural voice interaction reduces friction and saves time.
  • User experience: Smoother conversation handling may make AI voice agents more palatable to consumers, especially if interruptions are handled gracefully rather than awkward pauses.
  • Cost considerations: At $0.05 per minute plus backend costs, businesses must weigh improved experience against ongoing operational expenses.

Background

OpenAI's latest voice model advances on its Realtime API, launched in 2024, which also supported speech but was not full-duplex. GPT-Live-1 was initially released in July via the ChatGPT interface, giving users a more natural spoken interaction mode. The API release now allows developers outside OpenAI to embed similar capabilities. The model can operate alongside a backend language model (like GPT-6 Astra or a cheaper option) for tasks such as information lookup and tool use, separating conversational flow from computational heavy lifting.

Key Perspectives

OpenAI: Positions GPT-Live-1 as a major step forward for voice AI, touting benchmark performance and early customer results. Speak (language tutor): Reports an 80% drop in interruptions, suggesting the model gives learners a more natural cadence. Yelp (restaurant tech): Emphasizes improved guest experiences and revenue capture, though remains silent on whether diners accept or simply tolerate AI reservation agents.

What to Watch

  • Adoption rates: How quickly businesses move from pilot to full deployment, and whether customer satisfaction metrics improve.
  • Competitor response: Google, Amazon and others are developing similar voice AI — watch for announcements of rival full-duplex models.
  • Pricing evolution: Whether the $0.05/min rate drops as usage scales, and how custom voice pricing is structured for enterprises.

Sources

newspaper

Zotpaper

Articles published under the Zotpaper byline are synthesized from multiple source publications by our AI editor and reviewed by our editorial process. Each story combines reporting from credible outlets to give readers a balanced, comprehensive view.

Editorial ProcessSubscribe