OpenAI launches GPT-Live for ChatGPT Voice

OpenAI introduced GPT-Live on July 8, 2026, a new generation of voice models that now powers ChatGPT Voice. The company says GPT-Live is built for full-duplex conversation, which means it can listen and speak at the same time instead of waiting for neat turn-by-turn pauses. That may sound like a small interface upgrade, but it changes how people experience AI in practice: the assistant can wait while you think, handle interruptions more naturally and keep a conversation moving while harder work runs in the background.

The launch covers two models, GPT-Live-1 and GPT-Live-1 mini. OpenAI says they are rolling out globally across iOS, Android and ChatGPT.com. GPT-Live-1 becomes the default voice model for Go, Plus and Pro users, while the mini version becomes the default for Free users. API access is not available at launch, but OpenAI says developers and enterprises can sign up for updates.

Technically, the important shift is separation of roles. GPT-Live handles the live audio interaction, while more complex requests can be delegated to a frontier model behind the scenes. At launch, that background model is GPT-5.5. This means a user can ask for something that needs web search, reasoning or a longer answer without the voice experience freezing into silence.

For everyday users, the immediate value is lower friction. Voice becomes more useful for commuting, language practice, accessibility, quick brainstorming and hands-free help. OpenAI also says ChatGPT Voice can show visual cards for subjects such as weather, stocks and sports while the conversation continues.

For businesses and makers, the release points toward a more practical voice-agent layer. Customer support, training, field work and internal assistants all become more plausible when the AI can listen continuously, respond quickly and pass harder tasks to other models. The limitations still matter: OpenAI says GPT-Live does not yet support voice with video or screen sharing in ChatGPT, and some languages may have accent or fluency gaps.

The broader signal is that AI interfaces are moving beyond chat boxes. If GPT-Live works as described, speaking to an AI may start to feel less like dictation and more like collaboration.