GPT-Live-1 in the API

About GPT-Live-1 in the API

GPT-Live-1 in the API is an announced OpenAI developer offering for bringing the GPT-Live voice model family to API customers. OpenAI launched GPT-Live-1 and GPT-Live-1 mini in ChatGPT Voice on July 8, 2026 and stated that API access is planned. As of September 8, 2026, OpenAI's own public launch page still describes API availability as coming soon, so this Brandligo page remains an announced product line rather than claiming general availability.

What has OpenAI officially announced?

OpenAI introduced GPT-Live as a new generation of voice models designed for natural human-AI interaction. GPT-Live-1 powers paid ChatGPT Voice tiers and GPT-Live-1 mini powers Free users in supported regions.

The launch page also says OpenAI plans to bring the models to the API and provides a sign-up path for developers and enterprises. That is the current first-party basis for this record.

Is GPT-Live-1 generally available in the API yet?

No public OpenAI source opened for this review confirms general API availability for GPT-Live-1 itself as of September 8, 2026. OpenAI's launch page still says the models are planned for the API soon.

Developers should not design procurement, pricing or production timelines around an assumed release date. Brandligo will treat the line as announced until OpenAI publishes API documentation or a release notice that confirms availability.

What makes GPT-Live different?

OpenAI describes GPT-Live as full duplex, meaning it can listen and speak at the same time rather than waiting for rigid turn boundaries. It is designed to handle interruptions, pauses and more natural conversational flow.

For more complex questions, GPT-Live can delegate deeper reasoning or search to another frontier model while continuing the voice interaction. That architecture is aimed at more natural assistants than a simple speech-to-text, text-model, text-to-speech pipeline.

How should developers approach the current waiting period?

Teams that need production voice today should use the developer products OpenAI currently documents, including the Realtime API and its available realtime models. They should treat GPT-Live API access as a future migration or evaluation path rather than block a current project on an unannounced date.

A useful preparation step is to define latency, interruption handling, tool use, language coverage and cost targets now so GPT-Live can be benchmarked against the existing Realtime API when access becomes public.

What pricing information is available?

OpenAI has not published a general GPT-Live-1 API price on the source used for this review. Brandligo therefore does not infer pricing from ChatGPT subscriptions or from other realtime models.

When OpenAI releases API pricing, buyers should compare the full cost of audio input, audio output, tool calls and session duration rather than assuming a voice workload follows ordinary text token economics.

What are the main risks and unknowns?

The key unknowns are release timing, API pricing, exact model limits, supported endpoints and production rate limits. Voice quality can also vary by language, and OpenAI notes that some languages may have non-native accents or fluency gaps.

Voice agents can create higher operational risk when they are allowed to take actions. Developers should plan confirmation steps, scoped credentials and human review for consequential workflows.

Who should choose something else?

Use OpenAI Realtime API now if the application needs a supported production voice interface today. A specialist voice or contact-center platform may be better when telephony, routing, recordings or enterprise call operations dominate the use case.

GPT-Live API should be evaluated once OpenAI confirms general availability and publishes the commercial and technical details needed for a real buyer comparison.