What happened
In a November 5, 2026 announcement, Google introduced Gemini Omni 1.2 Flash as a generally available model for video creation and editing. The primary source says it is available through the Gemini API and Gemini for business, and describes Google Cloud as the infrastructure backing the release. Omni 1.2 Flash supports native video generation and editing up to 15 seconds, compared with a 10-second maximum for Omni 1.1. Users can specify a duration in one-second increments from 3 to 15 seconds.
Google lists the 720p price at $0.07 per second, which it says is 30% lower than Omni 1.1. The company’s suggested workflow is to draft at 360p or 720p, revise details through follow-up edits rather than regenerate the entire clip, and upscale selected work to 1080p or 4K. The announcement does not provide a measured comparison of total project costs; the proposed savings depend on the generation and revision process teams actually use.
A central feature is multi-turn conversational editing across full 15-second clips. Google says a user can make changes such as swapping a product or background, adjusting lighting or weather, and changing aspect ratio without starting over. It also describes guidance using character, product or style images, motion references up to three seconds, and audio cues. The company claims improved reference adherence and consistency for faces, garments, packaging typography, brand colors, objects and scenes across frames and camera transitions.
The announcement further claims more natural dialogue, clearer lip-sync, and improved multi-speaker conditioning, alongside more controllable continuous camera movement with fewer unwanted jump cuts. Google positions the model for developers, businesses and creators. It is rolling out across Flow, Google Vids, the Gemini app for Google AI Ultra subscribers, and Google Ads. The source says 15-second generation in Google Vids is available at no additional cost with a Google account or Google Workspace subscription; it does not give the subscription prices. Outputs carry an imperceptible SynthID watermark, and Provisioned Throughput is available. Google also flags fast sports action, rapid script changes and complex physics as areas still being improved.
What the archive adds
This is a new release in a developing Google video-generation thread, not a restatement of the earlier stories. Google’s agentic video understanding for Gemini Flash covered a different capability: analyzing long-form video while reducing processing costs. An earlier Telegram post on Gemini Omni 1.1 described its creative-control update; this announcement adds general availability, 15-second native clips, conversational editing and a lower stated API price.
Trend signal
This release supports the pattern that generative media is becoming application infrastructure: providers are offering generation through APIs with production-oriented controls and pricing. The archive’s assumption is that major providers keep making image, video and speech generation available as low-latency APIs with pinned versions. Google’s generative media becomes application infrastructure pattern fits the API access and per-second price, though this announcement does not establish latency or independently verify production performance.
Why now
Google presents this as its first Gemini Omni release to reach general availability, rather than a preview limited to selected users. The expansion coincides with a longer native clip limit and lower stated pricing, giving developers and businesses a defined API surface for production use. The source does not explain a competitive or technical trigger for the timing.
Why it matters
For teams producing ads, explainers, product demonstrations or other short videos, the most consequential change may be the combination of revisions and price rather than the 15-second limit alone. If a follow-up edit preserves the parts of a clip that already work, the team may avoid paying to regenerate an entire result. At $0.07 per second for 720p, the announced rate also gives teams a concrete starting point for estimating API generation costs.
The claims still need testing against a team’s own content, particularly where packaging text, identity, lip-sync or continuity matter. Production planning should include the costs of iteration and any upscaling, rather than treating the per-second rate as the full cost of a finished asset. General availability and Provisioned Throughput make this more actionable than a research demonstration, but they do not guarantee that every workflow will meet quality, reliability or budget requirements.
Limits and uncertainty
The source is Google’s product announcement and reports company claims and demonstrations; it does not provide an independent evaluation, controlled comparison or customer results. It does not specify generation latency, rate limits, regional availability, API quotas, or the separate cost of upscaling. The $0.07-per-second figure applies to 720p video, and the supplied material does not give equivalent prices for other resolutions. Google says fast sports action, rapid script changes and complex physics remain areas for improvement. The announcement’s examples of consistency and editing therefore should not be treated as guarantees for arbitrary prompts or production workloads.
Skill to build now
Run a small cost-and-quality trial before moving an existing video workflow. Select a representative shot, generate at 720p, then request one targeted change through a follow-up edit and compare that result with a fresh generation. Record clip duration, output quality, reference adherence and total generation spend, and separately check whether any upscaling is needed. This tests the specific claim that conversational revisions can reduce rework without assuming the stated per-second rate equals the cost of a finished asset.
Reader takeaway
Gemini Omni 1.2 Flash is now generally available for production use through Google’s API and cloud offering, with 15-second video and a stated 30% reduction in 720p pricing. The practical promise is that conversational edits and stronger reference consistency can make short-video workflows less wasteful. Validate that promise with your own clips before budgeting around it.