SMART WORLDJOURNALSubmit Manuscript
← All technology news

Artificial Intelligence

Google DeepMind Launches Gemini Omni 1.2 Flash with Enhanced Video Generation and Lower Costs

Google DeepMind’s latest AI video generation model, Gemini Omni 1.2 Flash, extends native clip length to 15 seconds, improves visual consistency, enables multi-turn conversational editing, and reduces costs by 30%, supporting scalable, high-quality professional content creation.

AI-ASSISTED REPORTING · NEWS

Google DeepMind has unveiled Gemini Omni 1.2 Flash, an advanced AI-driven video generation model now generally available via the Gemini API, Google Vids, and Gemini for Business. This update introduces native support for 15-second video clips, a 50% increase over its predecessor's 10-second limit, allowing creators to produce more complete stories and advertisements in a single generation.

The model features significant enhancements in reference adherence, maintaining consistency in faces, garments, typography, and brand colors across camera angles and transitions. This capability enables precise virtual try-ons, catalog spins, and branded video ads by faithfully preserving key visual elements throughout the clip.

Gemini Omni 1.2 Flash offers multi-turn conversational editing, permitting users to iteratively refine videos by adjusting elements like background, lighting, or aspect ratio without restarting the generation process. This streamlines workflows and facilitates efficient production of ad variants tailored for diverse audiences and placements.

Cost-effectiveness is a major improvement, with prices dropped by 30% from the previous version. Users can draft videos quickly in lower resolutions such as 360p or 720p, make targeted edits, then upscale the finalized clip to 1080p or 4K, optimizing resource use and expenditure for large-scale video production.

The technology also delivers better multi-speaker audio capabilities with improved lip-sync and natural dialogue flow. Cinematic controls allow smooth continuous camera movements, including flythroughs and zoom-outs, significantly reducing unwanted jump cuts and creating a polished visual experience.

Despite these advances, development continues to address challenges presented by fast-paced action, rapid script changes, and complex physics simulations. The current model suits professional settings for scalable, consistent video generation rather than spontaneous, highly dynamic scenes.

Google Cloud's enterprise infrastructure supports Gemini Omni 1.2 Flash, providing secure, reliable throughput for business applications. Early adopters include creatives, startups, and global brands integrating the model into diverse workflows across marketing and content production platforms.

Source & further reading

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

Source published 2026-11-05. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.