OpenAI's GPT-Live-1.1 Adds Cheaper Standby Pricing for Voice Agents
OpenAI released GPT-Live-1.1 to its API on 20 November 2026, a full-duplex voice model that lets callers interrupt it mid-speech. Active conversation is priced at $0.05 per minute, the same as its predecessor, while a new standby rate of $0.02 per minute applies when the line is open but silent. Standby pauses model inference while keeping the connection and history alive, and it is available only over WebRTC rather than SIP. A builder's worked example for a small-business receptionist estimated about $140 per 1,000 four-minute calls with half the time silent, versus $200 without standby. The author cautions that backend model and tool-call costs are billed separately and are not discounted by standby.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)
Log in to join the discussion and vote.
Log in