news
Google drops the paywall on AI video in Vids, running on Gemini Omni 1.1 Flash
Any Google or Workspace account can now generate 1080p scenes at no cost, with SynthID watermarking on every clip and no published free-tier quota.
Google has removed the subscription requirement for AI video generation in Google Vids. In a post dated September 23, 2026, the company said anyone with a Google or Google Workspace account can generate video at no cost using its Gemini Omni 1.1 Flash model, with output in 1080p. The workflow starts on desktop at vids.new, where users pick a "Create AI videos" option to begin scene generation.
Scene extension and exact clip duration are the two real controls
The generation controls Google described go slightly beyond typing a prompt and accepting whatever comes back. Users can set a specific duration for a generated clip, which matters when a scene has to sit under a fixed voiceover or beat. They can also extend an existing scene rather than regenerating it, with Google saying the extension keeps visual context, lighting, character appearance and environment consistent across the join. Existing AI clips can be upscaled, and new scenes can be created directly in 1080p.
Consistency across an extension is exactly the kind of claim that only holds up under repeated trials, and Google published no win rate, no comparison set and no failure rate alongside it. There is no stated maximum clip length in the announcement either, so the practical ceiling on a scene is unknown until people run it.
The free allowance is the number Google did not publish
Google's post points personal-account users who want more AI video generation toward its Google AI plans, and says Workspace Business and Enterprise plans include expanded generation pools plus centralized admin controls. What it does not say is how large the free pool is, how it resets, or what the paid pools contain. "Free" without a stated quota is a number-shaped hole, and it is the single figure that determines whether this is a usable production path or a demo allowance. Regional availability limits are also absent from the announcement.
The timing sits oddly next to the developer side of the same model family. The Gemini Omni line is expanding into a free consumer surface in the same month that the gemini-omni-flash-preview endpoint is scheduled for deprecation on September 30. Consumer access widens, API access churns.
SynthID on every clip, and a naming discrepancy worth flagging
Google said every clip generated with Omni 1.1 in Vids carries an imperceptible SynthID watermark embedded into the video frames, so viewers can verify the content was AI-generated. That is a provenance signal rather than a visible label, and it only helps anyone who has a SynthID detector in the loop.
Sources also disagree on which model is running. Google's own post names Gemini Omni 1.1 Flash. Android Authority's write-up refers to Gemini Omni 1.1 Flash-Lite being integrated into Vids while also saying 1080p generation runs on Omni 1.1 Flash. Flash and Flash-Lite are not the same tier, and the difference would show up in quality and speed, so the ambiguity is not cosmetic.
Text-to-speech is announced, not shipped
Google said Gemini 3.8 Flash-Lite text-to-speech is coming soon to Vids, turning a typed script into voiceover across more than 100 languages. No date was given. Until it lands, the free tier produces video and leaves narration to the user.
Sources: Google, Android Authority.
Daniel Ochoa: Covers model launches, shutdowns and pricing changes as they happen. Reads deprecation notices for a living so you do not have to.