Gemini Omni Flash arrives in Google Vids for Scheduled Release enterprise tenants on 5 August 2026

If your Google Workspace organisation uses Scheduled Release domains, 5 August 2026 is when Gemini Omni Flash begins rolling out to you in Google Vids. The gradual rollout takes up to 15 days, so not every user will see it on day one, but the majority of enterprise tenants who had been waiting since the Rapid Release domains got access on 16 July 2026 are now in scope.

There are three distinct things this update actually does, and they are worth separating out clearly.

Higher-quality AI video generation

The most straightforward change is that Gemini Omni Flash replaces the previous generation model inside Google Vids. The improvements Google cites are better text rendering, more realistic physics, and greater visual fidelity overall. If you have used Vids to generate short clips before and found the results unconvincing, the output quality has meaningfully improved.

Generated videos are 10 seconds long with native audio. Every clip includes an imperceptible SynthID digital watermark, which means AI-generated content is permanently identifiable even if the video is downloaded and shared outside Google’s ecosystem.

Conversational, text-prompt editing of any clip

This is the more significant capability shift. Gemini Omni Flash lets you edit existing video footage using plain-language instructions. You can ask it to fix the colour grading, restyle footage in a different visual treatment, or remove unwanted background audio such as traffic noise or an ambient siren. You type the instruction; the model applies the change.

What makes this different from earlier AI editing tools is the conversational architecture. Edits build on each other across multiple turns of a conversation rather than each instruction starting from scratch. Character consistency, lighting, and scene continuity are maintained across the session, so you can refine iteratively without the output drifting.

For anyone who produces regular video content for internal communications, training materials, or marketing, this significantly reduces the friction of repurposing existing footage. Changing the visual style of a clip for a different audience, or cleaning up audio from a recording taken in a noisy environment, no longer requires a separate editing tool or a video specialist.

One geographic note: editing non-AI video with Omni is not available at launch in the European Economic Area, Switzerland, the United Kingdom, Texas, or Illinois. If your team falls into one of those regions, the generation features are still available, but the edit-any-clip capability is restricted.

Personal avatar creation

The third feature lets you create a digital avatar of yourself that can be placed into AI-generated video clips. Setup takes under two minutes: a guided selfie and a 60-second voice sample. Once created, the avatar is stored in your Google Account and can be selected as a character within Vids.

The practical use case here is scaling your presence in video without booking studio time for every piece of content. An explainer video, a product walkthrough, or an internal announcement can feature you without requiring you to film it.

Avatars are available in English only, for users aged 18 and over, and are not available in the European Economic Area, Switzerland, or the United Kingdom at launch.

From an administrative perspective, personal avatars are on by default but can be disabled at the domain level through the Admin console. The video generation and editing features have no separate admin control.

Who has access and what it costs

Gemini Omni Flash in Vids is available to Google AI Pro and Ultra subscribers and Google Workspace business customers. Generation limits vary considerably: basic consumer accounts receive 10 Veo 3.1 generations per month, while Workspace AI Ultra subscribers can request up to 1,000 clips per month. Developers accessing the model directly through the Gemini API pay $0.10 per second of video output, putting a single 10-second clip at roughly one dollar.

Google Vids combines Veo 3.1 video generation, Lyria 3 soundtracks, and Gemini Omni editing in a single browser tab, which means you are not juggling separate tools for generation, music, and editing.

The broader picture

Gemini Omni Flash was announced at Google I/O on 19 May 2026 and made available to developers through the Gemini API and Google AI Studio on 30 June 2026. The same model powers avatar and remixing features in YouTube Shorts. Seeing the same capability distributed across Vids, Shorts, and developer tooling within roughly two months reflects a deliberate strategy of embedding one model across Google’s product lines rather than building separate AI features for each surface.

The competitive implication is straightforward. Services like HeyGen and Synthesia have built businesses around AI avatar video production for enterprise teams. Google is now offering comparable functionality as part of a Workspace subscription that organisations are already paying for. For many teams, the question of whether to adopt a standalone AI video tool just got simpler to answer.

If your organisation is on Scheduled Release domains, expect to see the Gemini Omni Flash features appear in Google Vids between 5 August and 20 August 2026. The official Workspace Updates blog has the full administrative detail, including the avatar governance settings available in the Admin console.