Google DeepMind Launches Gemini Omni 1.2 Flash with 15-Second Video Generation and Lower Pricing

Google DeepMind has announced the rollout of Gemini Omni 1.2 Flash, a major update to its video generation model that delivers native 15-second video creation, improved reference adherence, and multi-turn conversational editing [1]. The model is generally available across the Gemini API, Google Cloud, and enterprise business tools, carrying a 30% price reduction compared to its predecessor [1].

Enhanced Video Creation Capabilities

Gemini Omni 1.2 Flash extends native video generation limits up to 15 seconds, an increase from the 10-second limit found in Omni 1.1 [1]. Users can specify durations in one-second increments ranging from three to 15 seconds, allowing creators to produce complete stories, product explainers, and social advertisements in a single generation [1].

The model introduces multi-turn conversational editing, enabling users to refine clips turn-by-turn through dialogue rather than starting the entire generation process over from scratch [1]. This functionality permits adjustments such as swapping backgrounds, modifying lighting or weather, or changing aspect ratios across full 15-second clips [1]. Furthermore, the system supports fluid dialogue, clearer lip-sync, and multi-speaker conditioning for animated scenes and multi-character dialogue [1].

Improved Reference Adherence and Consistency

According to Google DeepMind, the updated model offers significantly better reference adherence [1]. Users can guide generation by combining style images, character or product photos, short motion clips of up to three seconds, and audio cues [1]. The model maintains consistency across faces, garments, packaging typography, and brand colors throughout camera angles and transitions [1].

This level of visual stability reduces morphing and visual drift from the first frame to the last frame, accommodating applications such as virtual try-on sequences, catalog spins, and branded video advertisements [1]. The system also enables continuous, complex camera movements—such as FPV drone flythroughs and dramatic zoom-outs—with fewer unwanted jump cuts [1].

Pricing and General Availability

Alongside the performance upgrades, Google DeepMind has lowered pricing by 30% compared to Omni 1.1, pricing 720p video generation at $0.07 per second [1]. The model is generally available for developers via the Gemini API and Google AI Studio, for enterprises through the Google Cloud console with Provisioned Throughput support, and across Google applications including Google Vids, Flow, the Gemini app, and Google Ads [1]. All generated outputs feature an imperceptible SynthID watermark [1].

Video generation models have rapidly evolved to offer longer context lengths, improved temporal consistency, and more precise prompt adherence, moving from short baseline loops to multi-second professional clips used in marketing and digital production environments.

For digital creators, startups, and enterprises in India and globally, lower pricing combined with native conversational editing and extended clip lengths lowers the barrier to scaling video production workflows, advertising campaigns, and multimedia content creation.

The release could accelerate adoption of generative AI video tools among marketing agencies and independent creators seeking cost-effective production pipelines. Enhanced consistency may reduce post-processing hours required to correct visual artifacts.

Google DeepMind notes that development is ongoing to further improve fast sports action, rapid script changes, and complex physics within the model.