OpenAI Adds a $0.02 Standby Rate as GPT-Live-1.1 Ships

OpenAI released GPT-Live-1.1 on its API on 19 November, its API changelog shows. The update adds a standby setting that pauses inference when a voice session goes quiet.

Standby minutes cost $0.02, against $0.05 for active minutes, per OpenAI’s model page. That is 60% less for each idle minute (our arithmetic).

Business Pill · The Model and the Harness

A short Business Engineer explainer on the difference between an AI model and the harness around it. It teaches the general idea only and does not discuss OpenAI’s pricing.

The key insight: OpenAI’s GPT-Live-1.1 keeps voice sessions at $0.05 per active minute and adds $0.02 per standby minute for paused WebRTC sessions, 60% less for idle time (OpenAI model page; our arithmetic).

What OpenAI Shipped

OpenAI describes GPT-Live-1.1 as a “Full-duplex voice model for multilingual conversations”. Full-duplex means it “can listen while it speaks”, the model page says.

Compared with GPT-Live-1, the page says, it “transcribes speech more accurately and maintains a more consistent accent and voice across languages”.

The changelog says developers can use gpt-live-1.1 in an existing Live session configuration. They describe in plain language when the model should delegate a task to their backend.

The cost of one hour of GPT-Live session time at OpenAI’s listed $0.05 active and $0.02 standby rates (o
The cost of one hour of GPT-Live session time at OpenAI’s listed $0.05 active and $0.02 standby rates (our arithmetic); backend model and tool use are billed separately.

How the Standby Rate Works

The changelog says developers can “Configure standby for GPT-Live-1 and GPT-Live-1.1 WebRTC sessions to pause inference during inactivity and resume when someone speaks”.

The model page lists two per-minute prices for live session duration: $0.05 active and $0.02 standby. “The lower standby rate applies to WebRTC sessions while inference is paused”, it says.

Standby is a WebRTC setting, per both pages. OpenAI does not list a standby rate for other connection types on the model page.

What It Means for an Hour of Voice

An hour fully active costs $3.00 at $0.05 a minute; an hour on standby costs $1.20 (our arithmetic). A session that is idle half the time comes to $2.10, 30% less than fully active (our arithmetic).

Backend model and tool usage is billed separately, the model page says, so these figures cover session time only, not the reasoning a voice agent hands off.

When GPT-Live-1 became generally available on 10 September, OpenAI’s changelog put voice sessions at “$0.05 per minute, billed per second”. The active rate is unchanged in the 1.1 listing. We compared GPT-Live-1 with xAI’s voice model in October.

The Structural Read

As we read the pricing, standby targets apps where the line stays open but no one is speaking, such as support calls on hold or always-on assistants.

As we read the release, OpenAI is pricing voice by the minute of attention rather than by the minute of connection.

OpenAI, GPT-Live-1.1 model page

“The lower standby rate applies to WebRTC sessions while inference is paused.”

Three Implications

FOR DEVELOPERS Standby can be configured on GPT-Live-1 and GPT-Live-1.1 WebRTC sessions, per OpenAI’s changelog.

FOR BUDGETS Session time runs $0.05 a minute active and $0.02 on standby; backend model and tool use are billed separately.

FOR MULTILINGUAL APPS OpenAI says GPT-Live-1.1 transcribes speech more accurately and keeps a more consistent accent across languages than GPT-Live-1.

What Is Not Established

OpenAI has not published accuracy figures for the transcription gain on these pages; the model page lists performance and speed as “Not specified”. How much of a typical session falls into standby depends on the app, so the savings above are examples, not averages.

This Week In The AI Supercycle

AI Lost Its Central Banker, the Lab Oligopoly, and the Future of Open Source

This week’s issue: the Nvidia leadership change, read as the AI build-out losing its central banker: who now finances the GPU landlords, backstops the data centres and sponsors open models, and why two labs are becoming the bank.

Read the full issue →

Business Engineer Framework

The AI Supercycle

The Business Engineer map of how AI capital, compute and pricing loop together.

Read the Map of AI →

The Bottom Line

GPT-Live-1.1 keeps the $0.05 active minute and adds a $0.02 standby minute for paused WebRTC sessions, per OpenAI, cutting the price of idle voice time by 60% (our arithmetic).

97,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.

A note on sourcing. Model capabilities, prices and the standby setting are from OpenAI’s API changelog (entries of 19 November and 10 September 2026) and its GPT-Live-1.1 model page, read on 19 November. The 60% and 30% figures and the hourly costs are our arithmetic on OpenAI’s listed per-minute prices.

Sources: OpenAI API changelog, 19 Nov and 10 Sep 2026 entries · OpenAI API, GPT-Live-1.1 model page

Scroll to Top
FourWeekMBA