Gemini Omni 1.2 Flash

Gemini Omni 1.2 Flash is Google DeepMind’s generally available video/omni generation model, announced on 2026-11-05. Blog: https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/ .

Positioning: a faster, lower-cost path for scalable video production, with longer clips, multi-turn conversational editing, and stronger reference adherence.

Highlights:

• Native generate/edit clips up to 15 seconds (up from 10 s on Omni 1.1); duration selectable in 1-second steps from 3–15 s;
• Multi-turn conversational editing: swap products/backgrounds, change lighting/weather or aspect ratio without regenerating from scratch;
• Stronger reference following: combine character, product, or style images, short motion clips (up to ~3 s), and audio cues;
• About 30% cheaper than Gemini Omni 1.1 — roughly $0.07/s at 720p;
• GA via Gemini API / Google AI Studio / Google Cloud; outputs carry an imperceptible SynthID watermark.

Problem

Ad spots, product explainers, and branded content need clips that are long enough, editable, and faithful to references. 1.2 Flash stretches single-shot generation to 15 seconds and uses conversational edits to cut full-regeneration cost.

Workflow & consistency

• Draft fast at 360p/720p, refine with follow-up turns, then upscale keepers to 1080p or 4K;
• Google reports stronger character/object/scene consistency from first to last frame with less morphing across motion and camera moves;
• Improvements called out for dialogue/lip-sync, multi-speaker conditioning, and continuous cinematic camera paths (FPV, zooms, orbits);
• Enterprise options include Provisioned Throughput on Google Cloud.

Limits

• Fast sports action, rapid script changes, and complex physics are still called out as ongoing work;
• Exact price tiers follow current Google docs/console;
• Consumer-app rollouts (Flow, Vids, Gemini app, Ads) may land on staggered schedules.

Who it's for

• Developers and marketing teams producing video ads, virtual try-on/catalog spins, and branded spots at scale;
• Teams on Gemini API, AI Studio, or Google Cloud;
• Creators editing multi-turn scenes inside Flow / Google Vids / Ads.

Getting started

1. Read https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/ ;
2. Developers: prototype in Google AI Studio, then Gemini API; businesses: try Google Cloud console;
3. Set 3–15 s duration, lock identity with references/motion/audio, and edit via conversation instead of full regen;
4. Expect SynthID on outputs; confirm regional pricing and quotas.