OpenAI Ships GPT-Live-1 Voice Model With Full-Duplex Telephony Support

OpenAI shipped a new full-duplex voice model in its API — meaning it can listen and talk at the same time instead of waiting for pauses — with built-in telephony support so developers can connect it directly to a phone number.

Key Points:

• GPT-Live-1 is a full-duplex voice model, meaning it can listen and speak simultaneously rather than waiting for the user to stop talking.

• The model includes built-in telephony support, allowing developers to connect it directly to a phone line for voice-based applications.

• Combined with the finance-specific ChatGPT release this week, this shows OpenAI increasingly building products shaped around specific channels and compliance needs rather than one general assistant.

Enterprise AI is splitting into narrow, regulated products — one voice model built specifically for phone-line conversations, one assistant built specifically for financial services — rather than a single chatbot trying to do everything.

Why It Matters: AI products are increasingly shaped around one channel's technical constraints — a voice model built specifically for phone conversations signals enterprise AI is going narrow and specialized.