Google announces Gemini 3.8 Live with Live Avatar for Gemini Enterprise
Google says Gemini 3.8 Live with Live Avatar is now available in Gemini Enterprise, combining live dialogue, streaming video, visual avatars, background tool calls, and multilingual speech-to-speech interaction. The announcement leaves major operational details unspecified, including plan and regional eligibility, pricing, quotas, supported tools, and custom-avatar approval criteria. Google says custom avatars require enterprise allowlisting and that AI-generated audio and video are marked with SynthID. [1]
This edition passed Imananq's enhanced publication checks. Some material claims remain explicitly attributed to official or company sources because no independent source is currently bound to this edition. The engine continues checking approved sources and will add corroboration only through a new edition that passes the full gate.
Google announced Gemini 3.8 Live with Live Avatar for Gemini Enterprise on September 24, 2026. The company says the feature combines live dialogue with low-latency streaming video, producing an audiovisual conversational persona. Its announcement also sets an important access boundary: custom avatars made from reference images are currently limited to enterprise allowlisting. [1]
01
What we know now
- 01
[1] Google, “Introducing Gemini 3.8 Live with Live Avatar,” 24 September 2026: https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-live-with-live-avatar
- 02
Google says the product is available in Gemini Enterprise; custom-avatar creation is limited to enterprise allowlisting.
- 03
Google says generated audio and video output is watermarked with SynthID.
02
Announced availability
Google says Gemini 3.8 Live with Live Avatar is available in Gemini Enterprise.Custom avatar access
Google says creating an avatar from a reference image currently requires enterprise allowlisting.Language claim
Google says the feature can transition across 97 languages; no language list or independent evaluation is provided.Output marking
Google says audio and video output from its AI products carries SynthID watermarking.Statuses and capability descriptions are from Google’s announcement and are not independently tested in the supplied evidence. [1]
03
A live audiovisual enterprise feature
Google’s September 24 post records the release of Gemini 3.8 Live with Live Avatar in Gemini Enterprise. According to Google, it pairs the company’s live-dialogue capabilities with low-latency streaming video to create a conversational agent with a visual presence. [1]
These are product descriptions from the developer, not independent findings. The supplied record provides no measured latency, reliability results, or account-level access information. [1]
- Google says the service can take visual and audio inputs at the same time and respond with expressive audio and video through a visual persona.
- The company describes lip-syncing, facial expressions, and turn-taking as part of the experience.
- Google presents customer service and interactive walkthroughs as possible uses, but the supplied evidence does not establish their performance or suitability in specific deployments.
04
Tool calls can run while a conversation continues
Google says Live Avatar supports asynchronous tool execution, intended to keep the conversation active while a background request is processed. This could matter for enterprise workflows that require information retrieval during a customer interaction, but the announcement alone does not show which tools or data sources are available. [1]
- Google says asynchronous tool calling lets the avatar initiate tool calls and data retrieval in the background while dialogue continues.
- A hotel check-in is an illustrative example in the announcement.
- The record does not identify supported tools, integration requirements, failure handling, or reliability limits.
05
Multilingual behaviour remains a vendor claim
The announcement makes a broad multilingual claim for Live Avatar. Readers considering multilingual use should distinguish the stated 97-language coverage from evidence of quality in particular language pairs, accents, or enterprise settings, which is not included in the supplied material. [1]
- Google says the feature uses native multilingual speech-to-speech synchronization and can transition across 97 languages.
- Google also says lip-sync and expressions adapt between languages without reduced video fidelity or visual drift.
- No language list, evaluation method, or independent assessment accompanies those assertions.
06
Custom avatars have a disclosed access restriction
For organisations seeking a branded or image-based avatar, allowlisting is the key practical constraint disclosed in the post. Google does not describe how an enterprise qualifies, what the approval process entails, or what consent and rights checks apply to reference images. [1]
- Google says organizations may select preset avatars.
- Google says developers can create an animated responsive avatar from a high-quality reference image while preserving likeness, styling, or character identity.
- Google states that custom-avatar creation is currently available only through enterprise allowlisting.
07
Watermarking is stated, but operational details are absent
Google presents SynthID watermarking as its transparency measure for AI-generated audio and video. Organisations that need provenance controls should seek the associated technical and policy documentation before treating this statement as a complete compliance or detection workflow. [1]
- Google says AI-generated audio and video output is watermarked with SynthID.
- It describes the watermark as imperceptible and intended to make generated material detectable.
- The supplied evidence does not specify detection procedures, coverage exceptions, or effectiveness evidence.
08
Primary source
This brief relies on Google’s announcement as the supplied complete primary evidence. It does not independently verify the company’s product-performance, safety, or coverage claims. [1]
- Primary source: Google, “Introducing Gemini 3.8 Live with Live Avatar,” published September 24, 2026.
- Link: https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-live-with-live-avatar
09
Before adopting it
The announcement confirms enterprise availability but leaves several operational details undisclosed. [1]
- 01
Confirm that your Gemini Enterprise account and region are eligible before planning a deployment. [1]
- 02
Request the current API documentation, pricing, quotas, supported tools, and technical requirements from Google. [1]
- 03
Treat custom-avatar access as restricted until enterprise allowlisting has been confirmed. [1]
- 04
Establish consent and rights checks for any reference image before submitting it; the announcement does not describe Google’s verification process. [1]
- 05
Test multilingual output, tool behaviour, latency, and watermark-detection workflows in the intended environment rather than relying on announcement claims. [1]
10
Limits of this edition
The supplied primary evidence does not identify eligible Gemini Enterprise plans, customers, countries, or regions. [1]
It does not provide pricing, quotas, API availability details, technical requirements, or a direct documentation record. [1]
Google’s capability, language-quality, latency, and safety descriptions are first-party claims; no independent testing is included. [1]
The announcement does not explain allowlisting criteria or consent, identity, and likeness protections for reference images. [1]
A linked model card and API documentation were not supplied as evidence, so their limitations and later changes cannot be assessed here. [1]
SRC
Source desk
Direct links to the material behind this selection. Seeing the source matters as much as reading the synthesis.
Keep exploring
3 Sept 2026
Google announces Gemini 3.8 Flash, with stated token pricing and higher-effort trade-off
Google announced Gemini 3.8 Flash with stated developer, enterprise and consumer access routes, plus introductory token prices matching those Google states for Gemini 3.7 Flash. Google says complex, higher-effort tasks can increase token use. Its separate Flash Cyber model is restricted through Fairwind, with key access terms unspecified. [1]
15 Sept 2026
Google announces Gemini 3.8 Live and Extended Thinking with staged access
Google announced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, describing live voice interaction, visual context and background task handling. The detailed access list places both models in the Gemini API and Google AI Studio, but enterprise access is private preview and some enterprise and Workspace releases are still described as coming soon. [1]
7 Sept 2026
Google announces agentic video understanding for Gemini API
Google says agentic video understanding is available through its Gemini API for three Flash models and can be enabled with an "agentic" processing setting. Google reports lower token use and costs and higher accuracy, and names LongVideoBench for one comparison, but does not provide the full benchmark methodology or conditions. It also says a Gemini app rollout will reach all users on Flash and Flash-Lite models soon, without a date or regional scope. [1]