OpenAI rolls out GPT-Live voice model worldwide
OpenAI announced global rollout of GPT-Live, a new voice-model architecture that can listen and speak at the same time and handle complex work in the background using GPT-5.5. The technology uses a full-duplex setup to allow brief acknowledgements, natural pauses or interruptions, and improved handling of background noise. Two variants will ship: GPT-Live-1 for paid tiers and GPT-Live-1 mini for free users, and the experience adds visual info cards for topics like weather and stocks. The company included voice-specific safety measures and limited voice impersonation, but video or screen-sharing voice calls are not yet supported.
GPT-Live listens and speaks simultaneously and delegates hard tasks to GPT-5.5
Context
Previous ChatGPT voice waited for users to finish speaking. Now OpenAI uses full-duplex voice that can process tasks in background. The company may add video or screen sharing support and future frontier model upgrades later.
The full analysis
19 dimensions on this story — world impact, market read, and what happens next.
- Full ContextLocked
- Affected SectorsLocked
- Stock ImpactLocked
- Economic IndicatorLocked
- Investor RelevanceLocked
- Professional RelevanceLocked
- Watch PointsLocked
- Probability of ChangeLocked
- Debate PointsLocked
- Prerequisite KnowledgeLocked
- Follow-up QuestionsLocked
- Pros & ConsLocked