GPT-Live: OpenAI's Full-Duplex Voice Model — Features, Access & Comparison
A current guide to GPT-Live's documented full-duplex voice behavior, consumer access boundaries, developer API path, and the evidence limits that matter when features or plans change.
This page separates what the current OpenAI Developer documentation supports from consumer ChatGPT details that can vary by account, surface, and region. The documented GPT-Live API is a full-duplex voice path that can listen and speak at the same time, with separate backend delegation. Consumer ChatGPT access and API model availability are different questions, so this guide does not infer a plan-to-model map from API pages.
What Is GPT-Live and Why Was It Released?
A Quick Look at OpenAI's New Voice Era
GPT-Live is described in the current Developer documentation as a full-duplex voice model: it can listen and speak at the same time and delegate reasoning or tool work to a backend. The public API documentation does not establish every consumer ChatGPT surface or replacement rule, so this page keeps those details separate.
Why GPT-Live Is a Major Leap in AI Conversation
The documented architecture separates the voice conversation from backend work. Responses delegation uses a supported Responses model; Client delegation lets an application run its own model, agent harness, or service. This is an API architecture fact and does not prove that every consumer feature, model name, or tool is available on every ChatGPT surface.
Key Features That Make GPT-Live Different
Full-Duplex Voice and Real-Time Interaction
The defining documented feature of GPT-Live is full-duplex conversation: it can hear audio while it is producing speech and can handle interruptions. The official docs do not provide a universal consumer timing number, so this page does not publish one.
More Human-Like Conversations
Natural backchanneling and interruption handling are product behaviors that can vary by model, prompt, account, and surface. The official API documentation supports the full-duplex and interruption model; exact consumer voice style should be checked against a current OpenAI product source.
Background Access to GPT-5.5 and Web Search
The current API docs describe delegation to a selected Responses model or to a client-run workflow. They do not establish GPT-5.5 as a universal backend or prove a consumer web-search behavior on every surface, so treat those details as unverified unless a current primary source supports the exact product claim.
Real-Time Translation and Smarter Dictation
The API model entry separates full-duplex voice from the gpt-live-transcribe transcription model. This page does not claim a fixed consumer translation or language-quality tier without a current primary source.
Visual Information Cards
Visual cards are a consumer-surface claim that was not verified in the current API evidence. Do not treat this feature description as proof that visual cards exist on every GPT-Live surface.
Upgraded Safety Mechanisms
The exact consumer safety presentation was not verified here. Apply the current OpenAI policies and product guidance, and keep application-level permissions and tool confirmation checks in your own integration.
GPT-Live Models and Availability
GPT-Live-1 vs GPT-Live-1 Mini
The current official API pages document gpt-live-1 for live voice and gpt-live-transcribe for transcription; they do not establish a gpt-live-1-mini API model. Consumer plan-to-model mapping requires current official consumer information and can vary by account or surface.
Consumer naming then moved on 15 October 2026, when OpenAI shipped GPT-Live-1.1 and a smaller GPT-Live-1.1 Quip to ChatGPT Voice, Codex and dots. Read what OpenAI shipped in GPT-Live-1.1 for the announced model split, the 13 new voices and the reported hour limits; none of that is an API model map, and this page does not treat it as one.
Supported Platforms and Rollout Schedule
Consumer availability can vary by account, app version, device, and region. This page does not claim a global rollout date or a fixed desktop schedule without a current official consumer source. Check the current ChatGPT product and Help Center information for your account.
When Will the GPT-Live API Be Available?
The API is documented now in the current OpenAI Developer pages. Start with the GPT-Live API guide, which covers gpt-live-1, v1/live/sessions, WebRTC, WebSockets, SIP, sideband, and delegation. This page does not claim a separate public waitlist form.
GPT-Live vs Advanced Voice Mode: What's New?
Turn-Based vs Full-Duplex Interaction
The API documentation supports a full-duplex model that can listen and speak at the same time and handle interruptions. Exact behavior of an older consumer surface requires current product documentation, so this comparison treats the consumer details as product-specific rather than universal API facts.
How Each Model Handles Interruptions
GPT-Live documentation describes interruption handling, but exact consumer pause behavior and account fallback behavior can change. Test the current surface you intend to support before presenting a comparison as a guarantee.
For a deeper side-by-side, see our full GPT-Live vs Advanced Voice Mode comparison.
GPT-Live is no longer the only frontier voice model in the conversation. SpaceXAI shipped Grok Voice 3.0 on 14 October 2026 with a reported first place on the Artificial Analysis speech-to-speech index — and the one support-call benchmark where GPT-Live-1 is still ahead. If you build on either API, the Grok Voice 3.0 prompt structure is a useful comparison for how much of an agent's behaviour lives in the system prompt rather than the model.
How to Enable GPT-Live in ChatGPT
Step-by-Step Setup Guide
- Open ChatGPT on web, iOS, or Android.
- Tap the voice icon (waveform) in the message input bar.
- If the current voice experience is available to your account, follow the on-screen voice controls.
- If the expected control is missing, update the app, check microphone permissions, and consult current OpenAI support information.
Free vs Paid: What You Actually Get
This page does not infer a current free-versus-paid model mapping from API documentation. Consumer entitlements can change by plan, account, surface, and region; use the GPT-Live pricing page as an editorial comparison and verify current plan details in official consumer information.
Looking for the live voice controls people mean when they search for ChatGPT live? The surface-by-surface walkthrough for web, iOS, and Android, including what the Live option does differently from Advanced and Standard, is in the enable guide.
Pros, Cons, and Known Limitations of GPT-Live
What works well
- Full-duplex voice is documented for the live API.
- Delegated backend work can be separated from the voice model.
- WebRTC, WebSockets, and SIP provide distinct connection paths for developers.
- Consumer feature quality and account access require current product evidence.
Where GPT-Live still falls short
- Consumer surface availability and feature parity vary by account and device.
- Language-quality tiers were not verified by a current primary source.
- API project access and rate limits must be checked before production use.
Frequently Asked Questions About GPT-Live
GPT-Live is described in the current Developer documentation as a full-duplex voice model that can listen and speak at the same time and delegate backend work. Consumer ChatGPT behavior is a separate evidence track.
Consumer access depends on the current plan and account. This guide does not publish a plan-to-model mapping without a current official consumer source, and API access is billed separately.
The current Developer documentation covers gpt-live-1, v1/live/sessions, WebRTC, WebSockets, SIP, sideband, and delegation. Confirm project access and limits before shipping.
Check the current official language support information for the surface and model you use. A supported-language list does not establish a GPT-Live voice-quality ranking; see the GPT-Live language guide for the evidence boundary.
Final Verdict: Is GPT-Live Worth Using in 2026?
GPT-Live is a useful label for a full-duplex voice API and a changing consumer product surface. Use the official API documentation for developer decisions, and verify account-specific ChatGPT access separately. For the interaction comparison, read our side-by-side comparison.
Sources
Citations used to fact-check this article. API architecture and pricing below were rechecked on 2026-09-19. Consumer plan and surface mapping is left unverified where no current claim-supporting source was captured.