GPT-Live-1: OpenAI's Full-Duplex Voice Model for Voice Agents in the API

Melvin Vivas · X video post · 2026-09-11 · 1:16 · 1,938 views · Open on X

Topics: AI Agents, Tool Use & MCP, LLM Fundamentals, Industry Trends & Job Market · Level: intermediate

Summary

This is a short launch demo of GPT-Live-1, a full-duplex voice model now in the OpenAI API. It can listen and speak at the same time, copes with background noise and interruptions, and hands tool calls and reasoning to a separate backend model. Melvin Vivas reshares the launch and says he wishes Codex could talk to users this way too.

Key points

• GPT-Live-1 is a full-duplex model: it listens and speaks at the same time, so users can interrupt it mid-reply.
• It is built for natural, expressive conversation and handles background noise and interruptions smoothly.
• Architecture: GPT-Live-1 is the front-end voice model, paired with a backend model that does the reasoning and tool calls. You can choose the backend model and agent harness.
• In the demo, the voice model hands actions (controlling a robot, updating a display) to the backend model while the conversation keeps going.
• It's described as 'ChatGPT Voice for your app': it brings ChatGPT Voice's natural back-and-forth into your own product.
• Pricing: $0.05 per minute for the front-end model. Backend inference and tool services are billed separately.
• Available in the API now and positioned for production voice agents at scale.

Resources mentioned

• GPT-Live-1 · tool · openai.com · paid
  OpenAI's full-duplex voice model in the API for real-time voice agents that pair with a backend reasoning/tool model.
  Also in: GPT-Live-1 Voice Agents in the API, and Charlie Guo (Melvin Vivas on X · notes)
• OpenAI API · tool · platform.openai.com · paid · recommended by both Bashiri Smith & Melvin Vivas
  OpenAI API documentation, including function calling and structured outputs.
  Also in: How to Relearn LLMs & RAG in 2026: A 7-Step Roadmap with Free Resources (Bashiri Smith on Facebook · notes), Livestream: Building a React Native ChatGPT App with Cursor and OpenAI (Melvin Vivas on X · notes), The 3 Levels of AI Engineering: LLM Apps → Production → Agentic Systems (Bashiri Smith on Facebook · notes), GPT Image 2 Adds Transparent Background Support in the OpenAI API (Melvin Vivas on X · notes) and 4 more
• ChatGPT Voice · tool · help.openai.com · free
  ChatGPT's voice conversation mode, used as the comparison point for GPT-Live-1's style of conversation.
  Also in: ChatGPT Voice Update: Plugins, Model Switching and ChatGPT Work (Melvin Vivas on X · notes)
• OpenAI Codex · tool · openai.com · paid
  OpenAI's coding agent. In the diagram it writes code, fixes review findings and drives the build loop. The creator also used it to make this video.
  Also in: An agent bot that installs and drives Codex on its own (Melvin Vivas on X · notes), Asking a Coder bot to install Codex (Melvin Vivas on X · notes), Sign in with ChatGPT: Setting Usage Limits for Each App (Melvin Vivas on X · notes), Codex Cloud Environments Must Be Saved & Published Before Use (Melvin Vivas on X · notes) and 240 more

Try this

• Try GPT-Live-1 in the OpenAI API to add conversational voice to your app.
• Pair GPT-Live-1 with a backend model of your choice for tool calls and reasoning.
• Plan costs as $0.05 per minute for the voice front end, plus separate backend inference and tool costs.
• Build a voice agent that hands actions (for example, controlling a robot or updating a display) to a backend model while the conversation keeps going, like the launch demo.
• Add a voice interface to a coding agent such as Codex (the creator's wish).