Gemini Omni 1.2 Flash: Price, Free Vids Access and Gemini App

Gemini Omni 1.2 Flash is Google's newest AI video-generation model, announced on November 5, 2026. It raises the maximum clip length to 15 seconds, cuts the API price to $0.07 per second of 720p video (down 30% from Omni 1.1), and powers a new free tier inside Google Vids that lets anyone with a Google account generate up to three 15-second AI videos a month at no cost. Here's what changed, what it costs, and how to actually use it.

What is Gemini Omni 1.2 Flash?

Gemini Omni 1.2 Flash is the latest version of Google's Omni line of AI video-generation models, built on the broader Gemini 4 model family. According to Google's announcement post, written by Anish Nangia, a product manager at Google DeepMind, the model is designed to be "fast and cost-effective" while supporting longer clips than its predecessor. It generates and edits video natively, meaning it creates motion, sound, and visuals together rather than animating a still image after the fact.

The headline change is clip length: Omni 1.2 Flash can produce videos up to 15 seconds long, compared with a 10-second ceiling on Omni 1.1. Google says duration can be set in 1-second increments starting at 3 seconds, giving creators finer control over pacing than a fixed-length clip would allow.

Omni is a distinct product line from Veo, Google's other text-to-video model family. Where Veo has historically been positioned as a standalone cinematic video generator, Omni is built to sit inside everyday productivity and consumer surfaces — Vids, Flow, the Gemini app — as well as being callable directly through the Gemini API for developers who want to embed video generation into their own products.

What's new compared with Omni 1.1 Flash?

Beyond the longer runtime, Google's blog post lists several specific improvements in Omni 1.2 Flash:

• Longer clips: Up to 15 seconds, versus 10 seconds previously, adjustable in one-second steps.
• Better reference adherence: The model more consistently matches characters, products, and visual styles supplied via reference images, short motion clips of up to three seconds, and audio cues.
• Stronger consistency: Less drift or morphing of characters, objects, and scenes from the first frame to the last.
• Multi-turn conversational editing: Users can refine a generated clip across several follow-up prompts — for example, swapping a background or product — without regenerating the whole video from scratch.
• More natural speech: Clearer lip-sync and distinct conditioning for multiple speakers in the same clip.
• More controllable camera movement: Fewer unwanted jump cuts when the model handles pans, zooms, or tracking shots.

Google is upfront that some limitations remain. The post specifically flags fast sports action, rapid script or scene changes, and complex physics as areas the team is still working to improve — a useful caveat for anyone judging the model against more specialized video tools.

Taken together, the changes read less like a ground-up rebuild and more like a focused iteration: Google kept the same underlying approach to native audio-visual generation and spent this release cycle on durability (longer clips without quality loss), steerability (better reference matching and multi-turn edits), and cost (the 30% price cut). That's consistent with how Google has iterated on its reasoning models too — frequent, incremental point releases rather than infrequent, sweeping overhauls.

Gemini Omni 1.2 Flash price: how much does it cost?

For developers using the Gemini API, Google AI Studio, or Vertex AI, Omni 1.2 Flash is billed by output tokens rather than a flat per-clip fee. At the standard rate of $12.00 per 1 million output tokens, and roughly 5,792 tokens generated per second of 720p video, the effective cost works out to about $0.07 per second — or roughly $1.04 for a full 15-second clip at 720p. That is a 30% reduction from Omni 1.1's per-second rate.

Pricing scales with resolution:

Google's pricing documentation also lists Gemini API input at $1.50 per 1 million tokens (text, image, or video) and text output at $9.00 per 1 million tokens for other Gemini models on the standard paid tier. Omni 1.2 Flash itself is generally available only on the paid tier of the Gemini API — it is not offered on the free API tier. Google's own guidance recommends drafting clips at 360p or 720p, using multi-turn edits instead of full regeneration, and upscaling to 1080p or 4K only once a cut is finalized, which keeps iteration costs down before paying for a higher-resolution render.

For context, a 15-second clip at 720p costing roughly $1.04 in API fees is a meaningful drop from what longer-form generative video has typically cost on a per-second basis industry-wide, though exact comparisons depend heavily on resolution, provider, and whether a request needs to be regenerated from scratch versus edited in place. Google's push toward multi-turn editing is partly a cost argument: an edit that only changes a background or product, rather than regenerating all 15 seconds, uses meaningfully fewer tokens.

Gemini Omni free: how Google Vids uses it at no cost

The same day Google introduced Omni 1.2 Flash, it published a second post announcing that Google Vids — Google's browser-based video creation and editing tool — now runs on the new model, with a free allowance attached. As the post puts it: "Starting today, anyone with a Google account or Google Workspace subscription can generate up to three 15-second videos per month in Vids at no cost."

That means the free tier is not unlimited, but it does not require a paid Gemini API key, an AI Ultra subscription, or a credit card — a standard Google account is enough. Google positions the feature around three use cases in the Vids post:

• Product promos: producing "professional, quick-turn social content and promotional materials without a single cut."
• Image-to-video: uploading reference images so characters and scenes stay consistent from the first frame to the last.
• Speech narration: generating clips with more natural-sounding narration.

Vids can be opened directly at vids.new. Google closes the announcement with the line: "If you can describe it, you can make it." For creators who have been comparing free AI video options, this sits alongside other generation tools covered on this desk, such as Sora 2 and Veo 3.1.

It's worth noting what the free allowance does not include. Google's post doesn't mention a resolution ceiling for the free Vids tier, doesn't say whether the three-videos-per-month cap resets on a calendar month or a rolling basis, and doesn't describe any paid upgrade path inside Vids itself for users who want more than three clips a month. Anyone who exhausts the free allowance and wants more volume would need to go through the metered Gemini API or Vertex AI instead, at the per-second rates above.

Gemini Omni video: where else can you use it?

Beyond Google Vids, Google lists several other ways to reach Omni 1.2 Flash:

• Gemini API: Generally available to developers, with full docs and pricing in Google AI Studio.
• Google Cloud / Vertex AI: Enterprise access through the Cloud console and API, including Provisioned Throughput for guaranteed capacity.
• Flow: Google's filmmaking tool, where creators can direct scenes and make multi-turn edits.
• Gemini app: 15-second video generation is rolling out to Google AI Ultra subscribers.
• Gemini for Business and Google Ads: Also listed as destinations for the model, aimed at business and advertising use cases.

Every video the model produces, regardless of where it's generated, carries an imperceptible SynthID watermark. SynthID is Google DeepMind's watermarking system, embedded at the moment of generation and designed to survive common edits like cropping, filters, and recompression, so AI-generated video can still be identified as synthetic later even after it has been downloaded, re-encoded, or re-uploaded to another platform.

Gemini Omni app: who gets free access vs. paid access?

It's worth separating the two tiers Google has described so far. Inside Google Vids, the 15-second, three-videos-per-month allowance is free to anyone with a Google account or a Workspace subscription — no AI Ultra plan required. Inside the Gemini app itself, 15-second video generation with Omni 1.2 Flash is rolling out specifically to Google AI Ultra subscribers, Google's top-tier consumer AI plan, rather than to all Gemini app users. Developers who want programmatic access, higher resolutions, or volume beyond the Vids free allowance need a paid Gemini API key through Google AI Studio or Vertex AI, billed at the per-second rates above.

Why Gemini Omni 1.2 Flash matters

The update continues a pattern of Google pushing AI video generation to be both faster and cheaper per generation rather than purely chasing higher benchmark scores. A 30% price cut alongside a 50% longer maximum clip length is a notable combination: historically, longer generative video clips have required more compute, not less cost per second. Bundling a free allowance directly into Google Vids also puts Omni 1.2 Flash in front of everyday Workspace and consumer users who might never open the Gemini API console, following the same distribution strategy Google has used with other AI features rolled into existing products.

It also arrives as Google continues to iterate rapidly across the Gemini 4 line, including the Gemini 4 Carbon model released in late October. Omni sits as a separate, specialized branch focused purely on video generation and editing rather than general-purpose reasoning, giving Google parallel release cadences for its text/reasoning models and its generative media models.

The Gemini brand itself, as Wikipedia's overview of the model family documents, has gone through several Flash-tier releases since Gemini 1.5 Flash in 2024, each one iterating on speed and cost rather than only raw capability. Folding a dedicated video-generation model into that same "Flash" naming convention — fast, affordable, high-volume — signals that Google wants Omni treated with the same expectations of speed and price-efficiency that Flash already carries for text and reasoning tasks, rather than as a slower, more expensive specialty product.

What's next for Gemini Omni

Google hasn't published a roadmap for the next Omni release, but the pattern so far — Omni 1.1 to Omni 1.2 Flash in a matter of weeks — suggests frequent iterative updates rather than infrequent major versions. Google's own post flags fast motion, rapid scene changes, and complex physics as open problems, which are reasonable areas to watch for the next update. For now, the practical takeaways are straightforward: developers get a meaningfully cheaper, longer-clip video model through the Gemini API and Vertex AI; everyday users get a no-cost way to generate longer AI video clips directly inside Google Vids; and Google AI Ultra subscribers get the same capability surfacing inside the Gemini app itself.

Written byTheo Park

Theo Park runs the AI desk at Pandromeda. He follows model launches from the frontier labs and the open-weight community, tracks the assistants and developer tools built on them, and explains what each release changes on pricing, capability and safety. His reporting leans on primary sources: model cards, technical reports, API documentation and the companies' own announcements.

More from Theo ParkAll AI coverage