What changed
On 19 November, OpenAI's API changelog announced GPT-Live-1.1 on the v1/live/sessions endpoint. Developers can describe in plain language when the model should delegate tasks to a backend. According to the changelog, standby mode is available for both GPT-Live-1 and GPT-Live-1.1 over WebRTC.1
How it works
OpenAI's model page describes GPT-Live-1.1 as a full-duplex voice model that listens while it speaks. It supports audio and text for both input and output, along with streaming and function calling. According to the model page, it does not support images, video, structured outputs, fine-tuning, or predicted outputs. Backend model and tool usage are billed separately from voice session duration.4
OpenAI's live conversations guide says GPT-Live-1.1 uses the same context events as GPT-Live-1. In a booking call, an application can send a lookup in progress through session.thinking.append so the model holds the fact without speaking it immediately. Once the backend confirms the booking, the app can send session.commentary.append so the model gives the caller the result. The guide also notes that GPT-Live summarizes older history in the background and starts a replacement voice engine when context usage exceeds 90%.3
According to OpenAI's model page, GPT-Live-1.1 and GPT-Live-1 have separate concurrent-session limits that apply across inference regions. The page does not specify performance or speed benchmarks for GPT-Live-1.1, and it does not list supported languages or general availability tiers. Developers are directed to their organization settings to check account-specific rate limits.4
Since publication
On 20 November, OpenAI's pricing page listed both gpt-live-1.1 and gpt-live-1 at $0.05 per minute for the active rate and $0.02 per minute for the standby rate, with voice sessions billed per second. That keeps pricing unchanged from GPT-Live-1 while letting WebRTC calls with long pauses drop to a lower rate during idle stretches. For comparison, the same page lists gpt-live-transcribe at an estimated $0.017 per minute and gpt-realtime-translate at $0.034 per minute.2
