Build4 publishers3 min readPublished Updated
Google ships Live Avatar for Gemini Enterprise agents before stating what the avatar costs
Google made Live Avatar, a lip-synced video face for Gemini 3.8 Live agents, generally available in Gemini Enterprise on September 24. Custom likenesses need an allowlist and Google has not disclosed an avatar rate, so teams can pilot preset avatars but cannot yet budget for them.
The Engineer · Build desk
Drafted by a language model from the sources cited here and checked against its claim ledger before publication. How we use AISend a correction
What happened
- Enterprises can pick from preset avatars now, while a custom avatar generated from a reference image requires Google's enterprise allowlisting.
- According to RuntimeWire, the launch materials link to a pricing page but state no Live Avatar-specific price or usage limits.
- Google Cloud's named customer is Cox Automotive, which built an avatar-based Autotrader car-shopping assistant, though Google gave no deployment numbers.
Compiled by The EngineerSomething wrong?How this is made
Why it matters
- decision A preset-avatar pilot can test whether customers complete tasks, but a production business case needs a per-session cost that stays unknown until Google states an avatar rate.
- exposure Companies building a custom likeness carry the face and voice consent risk themselves, so consent records have to exist before the avatar does.
- constraint Any roadmap built on a branded or real-person avatar waits on an approval process Google controls, so part of the launch date sits outside the team's hands.
Asynchronous tool calling is the part to test first. Google says Live Avatar can trigger tool calls and fetch data in the background while the dialogue continues [3]. In a blocking design, the agent goes quiet until the lookup returns. On video, the customer then watches an idle face. The asynchronous path hides the lookup behind speech without making it any faster, and that is still good engineering. Google's hotel check-in demo shows the agent calling tools for a guest while the conversation carries on [4].
The feature sits on a dialogue system Google shipped nine days earlier [18]. Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking came out on September 15 as models for real-time voice agents and task completion [16]. Live Avatar layers video onto that system [17]. The launch post was written by research scientist Shuo-yiin Chang and software engineer CJ Zheng of Google's Gemini Audio Team [5].
Google Cloud's documentation specifies 24-frame-per-second avatar video and sub-second interaction latency. RuntimeWire notes these are Google's specifications, not independent results [6]. Google also says lip-sync and expressions hold across 97 languages, including switches mid-conversation, without visual drift [8]. Treat both as measurements of Google's workload. For the latency figure to transfer, it would have to hold on the US or EU endpoint your users actually reach [9]. It would also have to hold on a turn with a tool call in flight and your own back end behind the tool.
Preset avatars are open to Gemini Enterprise customers. A custom avatar generated from a high-quality reference image is a separate path [10]. Google wrote: "Custom avatar creation is currently available only through enterprise allowlisting." [11] Google says all generated audio and video carries a SynthID watermark so it stays detectable as AI output [13]. Rights and consent for the face and voice samples behind a custom avatar are the customer's responsibility, according to Google's documentation [12].
Google did not disclose a Live Avatar rate or usage limits; its launch materials link to a pricing page, according to RuntimeWire [14]. A pilot can still answer the product question. Google Cloud's named customer example is Cox Automotive, which it says built an Autotrader assistant that uses an avatar to help shoppers search, compare vehicles and discuss financing [15]. A pilot built the same way, with a bounded task and a few tools, can show whether customers finish the task, though not what a session costs. Google pitches web, mobile and kiosk deployments [19], and finance will want a per-session number before any of them goes into production.
In my view, a team already running Gemini 3.8 Live voice agents should pilot a preset avatar now, because the avatar adds video to a dialogue system the team already operates [17]. I'd hold custom-likeness work until both the allowlist approval and the consent records are in hand [11][12].
What to watch
- Whether Google adds a Live Avatar rate and usage limits to the Gemini Enterprise pricing page.
- Independent latency measurements taken with tool calls in flight, or deployment figures from Cox Automotive's Autotrader assistant.
- Whether custom-avatar creation moves from enterprise allowlisting to general access.