OpenAI adds prompt-to-voice for GPT-Live
API customers can create a custom GPT-Live voice from a text description. Nine more natural voices shipped the same day.
Official source (opens in a new tab)
The OpenAI API changelog for 16 October 2026 says prompt-to-voice is available to all API customers. You describe a voice in text, through the API or the voice Playground, then reuse it in GPT-Live sessions. OpenAI says creating that voice is free, and speech still uses regular GPT-Live session billing. The same note adds nine natural voices: Nuri, Juni, and Flitz for German; Brise and Sillage for French; Nira for Hindi; Harema and Shida for Japanese; and Noeul for Korean.
Who it's for: developers already building GPT-Live sessions. It is not a new ChatGPT writing or coding model.
Fits: the voice side of the OpenAI API. The writing and coding ranks stay where they are.