OpenAI Launches GPT-Live: Full-Duplex Voice That Listens and Speaks at the Same Time

PrezenceAI Editorial · Published on 08/07/2026

🤖

This article was produced by PrezenceAI with the assistance of artificial intelligence. The main information was extracted from sources cited at the end of this text. The analysis, synthesis, and writing are the sole editorial responsibility of PrezenceAI.

🎧 Listen to this article

0:00 / --:--

The end of forced turns

On July 8, 2026, OpenAI launched GPT-Live, a family of voice models built on a full-duplex architecture — capable of listening and speaking at the same time. Two models rolled out globally: GPT-Live-1, the default for Go, Plus, and Pro plans; and GPT-Live-1 mini, the default for free accounts, replacing Advanced Voice Mode.

The architectural shift is fundamental. The previous system worked as a relay: transcribe speech, process through LLM, convert back to audio. Every turn required the user to stop speaking before a response arrived. GPT-Live processes continuously — making interaction decisions many times per second about when to speak, keep listening, pause, or invoke a tool.

Live translation and background delegation

The full-duplex architecture enables two use cases the old system could not deliver. First: real-time simultaneous translation — say "simultaneously translate what I am saying into Spanish" and the model begins translating while you are still speaking. At press briefings, OpenAI demoed the feature live and TechRadar called it "genuinely useful."

Second: background delegation — when a query requires web search, deep reasoning, or complex work, GPT-Live routes the task to GPT-5.5 in the background and keeps the voice conversation active. "This is exactly how humans interact with each other," said Atty Eleti, ChatGPT Voice product lead. "We keep the conversation going while we think in the background."

Availability and limitations

Launch covers iOS, Android, and ChatGPT.com. The desktop app and Business, Enterprise, and Edu workspaces were excluded at launch. No new pricing — GPT-Live is included in existing plans. OpenAI confirmed plans to bring the models to the API soon.

Declared limitations: GPT-Live is optimized for popular languages, with possible fluency gaps in others; does not support voice with video or screen sharing at launch.

The long-term bet

Over 150 million people use ChatGPT via Voice and Dictation, per OpenAI. The company sees voice as the future primary interface for complex long-horizon work — including agentic tasks that today belong to Codex. GPT-Live arrives on the same day as the broad GPT-5.6 rollout, signaling that July 8–9, 2026 marks a strategic inflection point in human-AI interaction.

Primary sources:
- OpenAI — Introducing GPT-Live: openai.com
- TechCrunch: techcrunch.com
- MacRumors: macrumors.com
- The Next Web: thenextweb.com