OpenAI Adds 13 Voices to GPT-Live-1.1 as ChatGPT Limits Vary by Plan

OpenAI began rolling out GPT-Live-1.1 in ChatGPT Voice on Oct. 15, adding 13 voices and plan-based usage limits to its real-time conversational feature across subscription tiers.

The six language groups each have two new choices: Aube and Rivage for French, Sonna and Miko for German, Leher and Neer for Hindi, Himari and Rindō for Japanese, Haetsal and Hanul for Korean, and Tinta and Arrullo for Mexican Spanish. Bubbie is the English option. Users can select voices in Settings under Voice, though available choices may differ by region and account.

Live can hear users while it speaks, allowing them to interrupt a response before it finishes. The update also gives Live more brief ways to signal that it is searching, reducing repetition of the same status phrase.

For individual plans, usage is measured over a rolling 24-hour period. Free accounts get limited access to GPT-Live-1.1 Quip, the smaller model, and the allowance may change. Go includes up to three hours of Quip, while Plus includes three hours of GPT-Live-1.1. The $100 Pro plan includes 15 hours; the $200 Pro plan offers unlimited use subject to abuse safeguards.

Business Standard includes three hours and Business Premium includes 15 hours. Use beyond those amounts costs 1.25 credits per minute. Enterprise usage billed in dollars costs $0.05 per minute.

Video and screen sharing are unavailable in Live. ChatGPT’s default personality settings also do not apply in Live, though users can ask it to change its tone or speaking speed. On mobile, users who need video or screen sharing can switch to the previous-generation Advanced Voice mode.

The update also reaches Codex and OpenAI’s always-on dots agents, extending the initial rollout of GPT-Live-1.1. OpenAI said API access is coming soon but gave no date.