Hands-free questions
Talk through everyday questions, planning, and explanations when typing is inconvenient.
Independent tool overview
GPT-Live at a glance
GPT-Live is OpenAI’s full-duplex voice system for ChatGPT, designed to listen and speak simultaneously, handle interruptions naturally, and delegate harder questions to frontier models.
Visit the official GPT-Live site ↗
- Product type
- Full-duplex conversational voice model in ChatGPT
- Models
- GPT-Live-1 on paid plans; GPT-Live-1 mini on Free
- Platforms
- Supported web, iOS, Android, and eligible workspace experiences
- Reasoning modes
- Instant, Medium, and High where the account supports them
- Maximum single session
- 2 hours
- Free access
- Limited GPT-Live-1 mini usage in each rolling 24-hour period
- Reviewed
- August 29, 2026
Overview
What GPT-Live is
GPT-Live powers ChatGPT’s newest Voice option. Unlike a turn-based assistant that waits for silence before responding, its full-duplex architecture continuously processes input while producing speech, letting it listen, pause, acknowledge, interrupt, or be interrupted during a conversation.
OpenAI offers GPT-Live-1 on paid consumer plans and GPT-Live-1 mini on Free. Users can choose Instant, Medium, or High intelligence where available; harder work can be delegated to a frontier model while the voice layer keeps the interaction moving. Voice can also use search and memory, accept text and images, and show visual cards for supported answers.
The improvement is conversational rather than infallible. Background noise, overlapping speakers, accents, network conditions, and timing can still cause interruptions or transcription errors. Important dates, locations, medical guidance, financial facts, and quotations should be checked in text or against an independent source.
Use cases
Who GPT-Live is best for
The strongest fit depends on the job you need the product to complete, not the size of its feature list.
Language and speaking practice
Run quick back-and-forth roleplay, ask for a different pace, and practice live translation while checking important phrasing.
Brainstorming aloud
Think through an idea naturally, interrupt the assistant, ask it to wait, and continue in the same chat with text or images.
Accessible multimodal help
Combine spoken questions with on-screen text, visual cards, typed messages, and attached images.
Voice-controlled work
Eligible desktop users can use Voice with Work or Codex to direct tasks and hear progress, under separate usage rules.
Capabilities
Core GPT-Live features
Full-duplex conversation
Listens and speaks at the same time so either side can interrupt, pause, or continue without rigid turn boundaries.
Active listening
Can acknowledge the speaker, wait through a pause, or stay quiet when asked rather than treating every silence as a completed turn.
Delegated reasoning
Hands harder search or reasoning tasks to a frontier model while maintaining the voice interaction.
Selectable intelligence
Instant favors responsiveness, while Medium and High spend more time on difficult questions when the plan supports them.
Search, memory, and visuals
Can use ChatGPT search and memory and show supported widgets for topics such as weather, markets, and sports.
Text and image input
Accepts typed messages and supported image attachments inside the same conversation while continuing to answer by voice.
Nine built-in voices
Offers Arbor, Breeze, Cove, Ember, Juniper, Maple, Sol, Spruce, and Vale, each with a different presentation style.
Background and CarPlay use
Can continue in the background on supported mobile devices and can start through Apple CarPlay on supported iPhones.
Audio provenance
Supported audio generated with GPT-Live includes SynthID watermarking, with OpenAI verification tools for detectable provenance signals.
Process
How the GPT-Live workflow works
Step 1
Select Live
Open ChatGPT Voice settings and choose Live if it is available for the plan, workspace, region, and app version.
Step 2
Choose voice and language
Set a preferred voice and primary spoken language to improve comfort and recognition.
Step 3
Set conversation behavior
Tell Live whether to answer quickly, wait until prompted, slow down, translate, roleplay, or use a more formal tone.
Step 4
Pick an intelligence level
Use Instant for fluid conversation and Medium or High when accuracy and deeper reasoning matter more than response speed.
Step 5
Speak and interrupt naturally
Correct a misunderstanding immediately, but move to a quieter environment if overlapping speech or noise becomes a problem.
Step 6
Review the record
Read the on-screen response and post-call transcript, remembering that the transcript is not a verbatim record.
Step 7
Manage stored clips
Delete the chat when its audio and transcript are no longer needed and review Data Controls before sharing recordings for model improvement.
Cost
GPT-Live pricing and free plan
GPT-Live is included with ChatGPT rather than sold separately. Free gets limited GPT-Live-1 mini access. Go and Plus each currently receive up to 1 hour of GPT-Live-1 Instant, 1 hour of Medium or High, and 2 hours of mini per rolling 24 hours. Pro offers much higher or unlimited access depending on the $100 or $200 tier.
ChatGPT Free
$0/month
Limited access to GPT-Live-1 mini.
- Usage measured over a rolling 24-hour period
- Exact mini limit can change
- Signed-in ChatGPT account required
- Single conversation can last up to 2 hours
ChatGPT Go
$8/month in the U.S.
Paid GPT-Live-1 access at the lowest consumer price.
- Up to 1 hour of Instant
- Up to 1 hour of Medium or High
- Up to 2 hours of GPT-Live-1 mini
- Limits reset on a rolling 24-hour basis
ChatGPT Plus
$20/month
The same published Live-hour allowances as Go, with broader Plus features and limits elsewhere in ChatGPT.
- Up to 1 hour of Instant
- Up to 1 hour of Medium or High
- Up to 2 hours of GPT-Live-1 mini
- Limits reset on a rolling 24-hour basis
ChatGPT Pro
$100/month tier
Substantially more voice time for frequent users.
- Up to 12 hours of Instant
- Up to 12 hours of Medium or High
- Up to 24 hours of GPT-Live-1 mini
- Single conversation still capped at 2 hours
ChatGPT Pro
$200/month tier
Unlimited GPT-Live-1 access subject to OpenAI’s guardrails.
- Unlimited GPT-Live-1
- Highest consumer access
- Single conversation can last up to 2 hours
- Usage remains subject to abuse guardrails
Pricing checked . Check current pricing at the source ↗
Assessment
GPT-Live strengths and limitations
Where it stands out
- Full-duplex interaction feels less rigid than turn-by-turn voice systems.
- Users can interrupt, pause, or ask the assistant to wait while thinking.
- Frontier-model delegation separates fast conversation from harder reasoning.
- Works across voice, text, images, search, memory, and supported visual cards.
- A limited mini model is available without a paid subscription.
- Multiple intelligence levels let users trade speed for deeper work.
- Audio clips are not used for model training unless the user opts in, according to OpenAI.
What to consider
- Live does not currently support video, screen sharing, connected apps, or plugins; Advanced Voice remains necessary for supported video and screen-sharing use.
- Live is not available inside custom GPT conversations, which continue to use Advanced Voice.
- Go and Plus each receive only one hour per rolling day at the selected full GPT-Live-1 intelligence category before other allowances or fallbacks apply.
- A single Live conversation is capped at two hours regardless of the plan.
- It is designed primarily for one speaker and can respond incorrectly when several people or background audio overlap.
- Transcripts may differ from what either side actually said and should not be treated as a verbatim recording.
- Some languages may have non-native accents or fluency gaps.
- Audio clips are stored with chat history for 30 days; deleting a chat triggers deletion within 30 days, subject to stated security, legal, and previously shared-training exceptions.
Compare
GPT-Live alternatives
The right alternative depends on the specific output, workflow, controls and budget your project requires.
Content Creator
Gemini
Compare Gemini for Google’s general-purpose assistant and its own live conversational experience.
Explore Gemini →Business Operations
Grok
Consider Grok for another consumer assistant with voice and real-time information access.
Explore Grok →Miscellaneous
GPT-Realtime-2
Choose OpenAI’s Realtime API model when building a custom voice application rather than using ChatGPT Voice.
Explore GPT-Realtime-2 →Content Creator
Voicenotes
Use Voicenotes when recording, transcription, and organized spoken notes matter more than a live assistant conversation.
Explore Voicenotes →Questions
GPT-Live FAQs
What is GPT-Live?
GPT-Live is OpenAI’s newest full-duplex voice system for ChatGPT. It can listen and speak simultaneously and delegate harder work to a frontier model.
Is GPT-Live free?
Free ChatGPT accounts receive limited GPT-Live-1 mini access. Full GPT-Live-1 is included with eligible paid plans under rolling daily limits.
How much GPT-Live time does Plus include?
OpenAI currently lists up to 1 hour with GPT-Live-1 Instant, 1 hour with Medium or High, and 2 hours with GPT-Live-1 mini in each rolling 24-hour period. Limits can change.
Can I interrupt GPT-Live while it is talking?
Yes. Full-duplex operation lets the system listen while speaking, although overlapping speech, noise, microphone quality, and network conditions can still cause mistakes.
Does GPT-Live support video or screen sharing?
Not currently. Eligible mobile subscribers can switch to Advanced Voice when they need supported video or screen sharing.
Can I use GPT-Live with a custom GPT?
No. OpenAI says Live is not available in custom GPT conversations; those use Advanced Voice instead.
Are GPT-Live transcripts exact?
No. The transcript may not exactly match either speaker, especially with overlap, background noise, or rapid conversation.
How long does OpenAI keep GPT-Live audio?
OpenAI says Live audio clips are stored with the chat transcript for 30 days. Deleting the chat leads to associated clip deletion within 30 days, with stated exceptions.
Does OpenAI train on GPT-Live audio?
OpenAI says it does not train on the audio clips unless the user chooses to share recordings for model improvement. Transcripts and other files may follow separate plan and Data Control settings.
Is GPT-Live available in the API?
OpenAI’s July 31 update says supported GPT-Live audio generated through ChatGPT Voice and the OpenAI API includes SynthID. Developers should verify the current model catalog, availability, and pricing before designing an integration.
Bottom line
Our GPT-Live verdict
GPT-Live is a meaningful usability upgrade for people who already talk to ChatGPT: interruptions, listening behavior, selectable reasoning, and multimodal chat make it much easier to think aloud. Free is enough to test the experience; Go or Plus suits occasional use, while Pro pricing only makes sense for people who genuinely need hours of voice every day. Keep the text view open for consequential answers, and treat the stored transcript as a convenience rather than an exact record.
Visit GPT-Live website ↗Similar Consumer Tools
View all Consumer AI tools →
Grok 4.5
Grok 4.5 - SpaceXAI and Cursor's strongest Grok model with powerful performance and efficiency at low costs
Muse Spark 1.1
Muse Spark 1.1 - Meta's upgraded, cost-effective agentic model with 1M-token memory and computer use
Willow Frontier Mini
Willow Frontier Mini - Willow's free, unlimited AI dictation model
ChatGPT Work
ChatGPT Work- OpenAI's Codex-powered agent for everyday non-coding work
AI training for the future of work.
Get access to all our AI courses, hundreds of real-world AI use cases, live expert-led workshops, an exclusive network of AI early adopters, and more.
AI Courses
Get unlimited access to all of our current & upcoming industry-specific AI courses for the duration of your subscription.
Daily Guides
To keep up with the rapid pace of AI, our team publishes AI implementation guides daily. Our library contains 300+ practical use cases to automate real-world work.
Workshops
Join weekly, live, interactive sessions with industry leaders who are at the forefront of AI for hands-on implementation guidance and exclusive insights.
Community
Network with an exclusive community of AI-first professionals who are working smarter with AI. Learn how early adopters are using AI in their work and businesses.