# Google introduces Gemini 3.8 Live with Live Avatar

- Source: MechNext — https://mechnext.eu/en/article/20-google-introduces-gemini-3-8-live-with-live-avatar
- Published: 2026-09-25
- Updated: 2026-09-25
- Category: Industry News
- Language: en (Markdown in other languages: ?lang=de, ?lang=sk, ?lang=pl, ?lang=zh, ?lang=ru)

> Gemini 3.8 Live with Live Avatar is available in Gemini Enterprise. Google describes near real-time video paired with speech, synchronization across 97 languages, and an enterprise allowlisting requirement for custom avatars.

In an announcement dated September 24, 2026, Google introduced Gemini 3.8 Live with Live Avatar. The company says it is available in Gemini Enterprise. Google describes Live Avatar as combining near real-time video generation with speech, presenting the visual and spoken parts as a coordinated experience rather than as separate outputs.

### Video, speech, and turn-taking

According to the announcement, Live Avatar is designed to pair generated video with speech. Google highlights precise lip-syncing, natural expressions, and fluid turn-taking as characteristics of the experience. Those details describe how the avatar is intended to accompany dialogue: spoken output is paired with facial movement and expression, while the conversation can move between turns. The announcement does not provide measurements or comparative results for these characteristics, so they remain the company’s description of the product.

Google also says Live Avatar can make asynchronous tool calls and fetch data in the background while dialogue continues. In the announced design, background retrieval can happen alongside an ongoing exchange rather than requiring the dialogue to stop first. The source does not identify particular tools, data sources, or use cases, so the capability is best understood at the level stated: asynchronous calls and background data fetching during continuing dialogue.

### Synchronization across languages

The announcement says Live Avatar supports speech-to-speech synchronization across 97 languages. Google adds that lip-sync and expressions adapt across languages. This makes language support part of the described audio and video coordination: the statement covers synchronization and visual adaptation, not just the availability of speech in multiple languages. The source does not list the 97 languages or describe differences in support among them.

For organizations considering the feature, the wording establishes the scope of Google’s claim without specifying how language selection works or how the adaptation is implemented. It says that synchronization spans 97 languages and that lip movements and expressions adapt across them. No additional performance figures or language-by-language assessments are provided in the announcement.

### Custom avatars for organizations

Google says organizations can customize Live Avatars using a high-quality reference image. The image is used to generate an animated, responsive avatar. This describes an organization-facing customization option based on a reference image; the announcement does not give further details about the creation workflow, image requirements, or how the resulting avatar is deployed.

Custom avatar creation is currently available only through enterprise allowlisting, according to the source. That access condition is relevant to organizations evaluating customization: the announcement does not describe custom creation as generally open to every Gemini Enterprise user. It specifically states that enterprise allowlisting is required at this time. The source gives no further information about eligibility criteria or the process for requesting access.

### Watermarking and getting started

Google says AI-generated audio and video output is watermarked with SynthID. The statement applies to both modalities named in the announcement. It does not explain how the watermark works or how it can be detected, so the supported description is simply that SynthID is used to watermark generated audio and video.

For those looking to get started with Gemini 3.8 Live with Live Avatar, Google provides API documentation. The announcement points to that documentation as a starting path, without detailing integration steps or technical requirements in the article itself. Together, the stated availability in Gemini Enterprise, the combination of video and speech, background tool calls, language synchronization, optional custom avatar creation subject to allowlisting, and SynthID watermarking make up the product details provided in the announcement.

These details come from Google’s announcement, “Introducing Gemini 3.8 Live with Live Avatar,” dated September 24, 2026. They reflect the company’s description of the product and its stated access conditions. The announcement is the basis for the reported capabilities and availability here.