While video generation under Omni 1.1 was limited to 10-second outputs at higher rates, Google has launched Gemini Omni 1.2 Flash to generate and edit native 15-second clips at $0.07 per second for 720p video. As Google stated, the release marks a 30% price reduction from Gemini Omni 1.1, making the model generally available across the Gemini API and Gemini for business.
This pricing drop and the introduction of iterative conversational editing shift AI video production from an expensive testing ground into regular operational workflows. For enterprise ad production budgets, this reduction directly alters the unit economics of high-frequency asset testing. Traditional video pipelines require significant capital allocation for exploratory takes and minor adjustments, whereas a $0.07 per second price point for 720p generation allows marketing teams to scale up variant output without expanding initial media spend. The commercial stakes center on whether teams can maintain brand consistency across multi-shot sequences without inflating manual post-production budgets, especially when balancing lower compute expenses against the labor required for artifact cleanup.
When evaluating the total cost of implementing the new generation model versus legacy workflows, decision-makers must weigh the immediate compute savings against the hidden overhead of human oversight. Legacy workflows often incurred high costs through iterative full-clip re-render cycles, whereas Gemini Omni 1.2 Flash reduces wasted compute by enabling targeted adjustments. However, if lower compute costs are counterbalanced by the labor required for post-production artifact cleanup—such as correcting frame inconsistencies or edge flickering—the net operational savings narrow. In our view, the financial advantage is realized only when the model's reference-locking features successfully minimize manual touch-ups.
Multi-Turn Workflows and Layered Editing
Under earlier iterations, modifying a scene meant re-generating the entire clip from scratch and paying full compute costs for each iteration. In Gemini Omni 1.2 Flash, conversational multi-turn prompts let users alter specific elements turn by turn across full 15-second clips without discarding previous visual work.
"We have dropped the price 30% from Gemini Omni 1.1 making it available at $0.07 per second for 720p video."
Production teams can draft quickly at 360p or 720p, apply a follow-up edit to adjust details, and upscale final creations to 1080p or 4K. Creators can set clip durations in one-second increments ranging from 3 to 15 seconds. For background adjustments, conversational editing allows environments to swap across locations while preserving character performances, camera framing, and surface reflections, effectively reducing the compute waste typical of legacy pipelines.
Brand Adherence and Enterprise Deployment
A central operational hurdle in automated video creation is preventing brand assets and character identities from drifting across frames. Gemini Omni 1.2 Flash addresses this by retaining references from start to finish, accepting inputs that combine character, product, or style images, short motion clips up to three seconds, and audio cues.
Multi-turn conversational editing allows refining the same clip turn by turn without re-briefing or re-generating the entire video.
This reference locking ensures that packaging typography, garment silhouettes, and brand colors remain stable across camera movements, which directly impacts the labor required for post-production artifact cleanup. The model can combine a single model photo with front-and-back product shots of three outfits to generate a 15-second virtual try-on sequence across 360-degree rotations and match cuts while maintaining garment drape and identity details. By keeping these elements consistent without requiring manual re-renders, the system helps offset the operational friction that typically accompanies automated video production.
The catch
While video generation used to be an expensive experiment plagued by paying for discarded full reruns, Google Gemini is now turning this pipeline into routine operations. The company systematically pushes aggressive unit economics across its lineup, cutting wasted compute right inside the conversational prompt interface. Yet slashing generation fees immediately exposes the true operational friction — the cost of human labor spent scrubbing visual glitches and frame-to-frame inconsistencies.
Our stories: Google's Video AI Economics: Nano Banana and Gemini Omni (July 1).