OpenAI · 2026-07-08 · major
OpenAI GPT-Live — full-duplex voice model now powers ChatGPT Voice
OpenAI launched GPT-Live, a full-duplex voice model family that becomes the default ChatGPT Voice on iOS, Android, and the web. GPT-Live-1 powers Go, Plus, and Pro; GPT-Live-1 mini is the default for Free users.
OpenAI GPT-Live is a full-duplex voice model that lets ChatGPT listen and speak at the same time.
Quick facts
| Maker | OpenAI |
|---|---|
| Models | GPT-Live-1, GPT-Live-1 mini, GPT-Live-1 Medium, GPT-Live-1 High |
| Architecture | Full-duplex (listens and speaks at the same time) |
| Availability | ChatGPT iOS, Android, ChatGPT.com — rolling out globally starting July 8, 2026 |
| Default for | Go, Plus, Pro (GPT-Live-1); Free (GPT-Live-1 mini) |
| Reasoning handoff | GPT-5.5 in the background for search, deeper reasoning, and complex work |
| Not yet supported | Voice with video or screen sharing |
What is it?
GPT-Live is OpenAI's new generation of voice models built for continuous conversation instead of turn-by-turn exchanges. GPT-Live-1 replaces Advanced Voice Mode for Go, Plus, and Pro subscribers; GPT-Live-1 mini is the default for Free users; and Medium and High tiers route to GPT-5.5 Thinking for harder work.
How does it work?
Instead of processing a sequence of separate messages, the model handles audio input and output simultaneously and makes interaction decisions many times per second — whether to speak, keep listening, pause, interrupt, or invoke a tool. Complex questions get handed to GPT-5.5 in the background, and GPT-Live keeps the conversation going until the frontier model returns an answer.
Why does it matter?
Around 150 million people use ChatGPT Voice every week, and turn-taking has been the main friction — the previous voice mode would interrupt users mid-thought or fail to hold context across a long conversation. GPT-Live lets people interrupt, pause, or ask ChatGPT to slow down, so voice becomes usable for real, long-running tasks rather than only quick queries.
Who is it for?
ChatGPT users on iOS, Android, and web; voice-agent teams tracking OpenAI's default voice stack.
Frequently asked questions
- How does GPT-Live differ from ChatGPT Advanced Voice Mode?
- GPT-Live uses a full-duplex architecture so ChatGPT can listen and speak at the same time, backchannel with phrases like 'mhmm', wait when a user pauses to think, and delegate heavier questions to a frontier model in the background — instead of Advanced Voice Mode's turn-based back-and-forth.
- Who gets GPT-Live-1 versus GPT-Live-1 mini?
- GPT-Live-1 becomes the default ChatGPT Voice for Go, Plus, and Pro subscribers, while GPT-Live-1 mini is the default for Free users. Both share the same continuous-interaction interface, and Medium and High tiers route the background reasoning to GPT-5.5 Thinking for harder questions.
- Does GPT-Live support voice with video or screen sharing?
- At launch GPT-Live does not support voice with video or screen sharing — users who need those features can fall back to the previous ChatGPT Voice experience. OpenAI is shipping visual response cards for topics like weather, stocks, and sports as GPT-Live's launch visual surface.
- What frontier model does GPT-Live call in the background?
- At launch GPT-Live delegates to GPT-5.5 when a question needs web search, deeper reasoning, or more agentic work, and it keeps the conversation flowing while that runs. OpenAI says the background model will be updated as newer frontier models become available.
Try it
Tap the Voice button in ChatGPT on iOS, Android, or ChatGPT.com — GPT-Live is the new default.