Google’s Gemini Omni 1.2 Flash makes 15‑second AI video cheaper and easier to edit

Want to spin up a short product video without hiring a crew or paying for expensive rendering? Until now, Google’s generative‑video model capped clips at ten seconds, and any tweak meant re‑generating the whole sequence, which quickly ate into budgets and timelines, especially for small teams that need to iterate quickly on branding or messaging, or for freelancers juggling multiple clients.

Google’s new Gemini Omni 1.2 Flash changes that workflow. It generates native video up to fifteen seconds and does so at a lower per‑second price, letting creators produce longer snippets and adjust details without starting from scratch. This opens the door for quick product demos, social‑media ads, or educational explainers that fit within typical platform limits and keeps the overall cost manageable for tight marketing budgets.

Source: Google Gemini Blog: Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips ↗

Google’s blog confirms that Omni 1.2 Flash supports native video generation up to fifteen seconds, a jump from the ten‑second cap of the prior Omni 1.1 release. The announcement also notes a 30 % price cut, pricing the model at $0.07 per second for 720p output. This pricing makes short‑form video generation affordable for small businesses and independent creators who previously faced high per‑frame costs.

The update also adds multi‑turn conversational editing, so users can ask the model to swap a background, change lighting, or replace an object within the same clip while the system retains reference fidelity across frames. Google says the model is generally available through the Gemini API, Google Cloud and integrated Google apps. This capability reduces the need to re‑render entire videos when only minor tweaks are required, speeding up creative iteration.

Source: Google Gemini Blog: Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips ↗

Despite the improvements, the model remains limited to fifteen‑second clips, so longer narratives still require stitching multiple generations or using traditional tools. The announced $0.07 per‑second rate applies to 720p video; pricing for higher‑resolution outputs was not disclosed, and Google labels the technology as experimental, indicating that features like ultra‑smooth motion or complex physics are still under development.

Reference adherence is stronger than in Omni 1.1, but the blog notes that visual drift can still occur on very intricate scenes, meaning fine details such as tiny typography or subtle lighting shifts may not always stay perfectly consistent across frames. Creators should plan for occasional manual touch‑ups when precision is critical, especially for brand assets that demand exact color or logo placement.

Developers can start by signing up for the Gemini API and exploring the starter apps in Google AI Studio, which include sample code for generating fifteen‑second clips and invoking multi‑turn edits via simple prompts. The API’s pricing page shows the $0.07‑per‑second rate, so you can estimate costs before building a production pipeline. Run a quick test with a product image and a brief script to see how the model handles reference consistency.

Marketers can try the built‑in Flow editor or Google Vids inside a Workspace account to create ad variants without extra cost, using conversational prompts to swap backgrounds or adjust lighting on the fly. Keep an eye on Google’s roadmap for higher‑resolution pricing and future enhancements like faster sports motion, then decide whether to integrate the model into larger campaigns.