ChatGPT✓ ReviewedAnnounced 8 July 2026Reviewed 18 October 2026GPT-Live listens while it talks. Fast answers still deserve slow checks.
Reviewed editorial Update
GPT-Live listens while it talks. Fast answers still deserve slow checks.
OpenAI's GPT-Live voice models hold full-duplex conversation, listening and speaking at once, and quietly hand harder questions to a stronger model in the background across ChatGPT web, iOS and Android.
What changed
OpenAI introduced GPT-Live-1 and GPT-Live-1 mini, full-duplex voice models that can listen and speak simultaneously. The current Chat voice experience has since moved to GPT-Live-1.1 and GPT-Live-1.1 Quip: Free has limited Quip access, Go uses Quip and paid tiers include Live-1.1 subject to their own allowances. Live can use connected apps, while video and screen sharing remain in Advanced on supported mobile accounts. In the API, developers choose the backend model or agent independently of the voice model; OpenAI's API launch article names GPT-6 Astra as one example. Supported generated audio also carries SynthID watermarking under OpenAI's provenance system.
Why it matters
Conversation that flows like a person removes the pause that used to remind you a model was answering. The interface got more convincing, not more correct, and a confident spoken answer is harder to fact-check mid-flow than a paragraph you can reread.
Bit’s takeaway
The human take
The person keeps the habit the interface no longer prompts: pausing to verify before a fluent answer becomes a decision.
Turn this into a workflow
Bring your task to Bit.
Bit reevaluates the recorded facts when you ask. Historical recommendations do not approve a current action.
Ask Bit about my taskSource-backed facts
- OpenAI introduced GPT-Live-1 and GPT-Live-1 mini for simultaneous listening and speaking. Its current Chat Voice documentation now names GPT-Live-1.1 and GPT-Live-1.1 Quip: Free has limited Quip access, Go uses Quip, and eligible Plus, Pro and Business seats use GPT-Live-1.1. In the API, GPT-Live can delegate task reasoning to a separately chosen model or agent. Chat, desktop Voice and API usage have different limits and billing.Reviewed
Technical record
[openai-gpt-live-fact; official_fact; product_fact; reviewed 2026-10-18; sources: OpenAI: ChatGPT Voice (https://help.openai.com/en/articles/20001274-chatgpt-voice); OpenAI: GPT-Live API guide (https://developers.openai.com/api/docs/guides/live); OpenAI: Introducing GPT-Live-1 in the API (https://openai.com/index/introducing-gpt-live-1-in-the-api/)] OpenAI introduced GPT-Live-1 and GPT-Live-1 mini for simultaneous listening and speaking. Its current Chat Voice documentation now names GPT-Live-1.1 and GPT-Live-1.1 Quip: Free has limited Quip access, Go uses Quip, and eligible Plus, Pro and Business seats use GPT-Live-1.1. In the API, GPT-Live can delegate task reasoning to a separately chosen model or agent. Chat, desktop Voice and API usage have different limits and billing.
- OpenAI added SynthID to supported audio generated through ChatGPT and the OpenAI API, and expanded public and API verification to supported audio.Reviewed
Technical record
[openai-audio-provenance-fact; official_fact; product_fact; reviewed 2026-10-10; sources: OpenAI content provenance and verification (https://openai.com/index/advancing-content-provenance/)] OpenAI added SynthID to supported audio generated through ChatGPT and the OpenAI API, and expanded public and API verification to supported audio.
Sources and method