Introduction

On November 5, 2026, Google rolled out Gemini Omni 1.2 Flash and made Gemini Omni generally available on the Gemini API and Google Cloud. The maker-facing upgrades are simple to pin: native clips up to 15 seconds (was 10s on Omni 1.1), multi-turn conversational editing, stronger reference adherence, and a 30% price cut to $0.07 per second at 720p.

This is the production GA for the Omni Flash video lane — SynthID watermarked outputs, Provisioned Throughput on Cloud, and day-0 surfaces in Flow, Vids, Ads, and the Gemini app (Ultra rollout).

What shipped

Concrete facts from Google’s developer blog and the same-day Cloud post:

  • Product: Gemini Omni 1.2 Flash (video generation + edit)
  • Vendor: Google DeepMind / Google (closed weights / API)
  • Modality: text-to-video, image-to-video, video edit; native audio / multi-speaker dialogue (audio references still allowlisted on Cloud)
  • Duration: generate and edit 3–15 seconds in one-second steps (up from 10s on Omni 1.1)
  • Price: $0.07/s at 720p — ~30% less than Omni 1.1 Flash; similar savings at other resolutions
  • Workflow tip: draft at 360p or 720p, then upscale keepers to 1080p or 4K
  • Enterprise: first GA for Gemini Omni on Gemini API + Google Cloud; Provisioned Throughput; imperceptible SynthID watermark

What actually got better

Beyond the longer window and the discount:

  • Reference adherence — combine character / product / style images, short motion clips (up to ~3s), and audio cues; faces, garments, packaging type, and brand colors hold across angles
  • Consistency — less drift across camera moves, first-to-last frame paths, and clip extensions
  • Multi-turn edit — swap a product or background, change lighting or aspect ratio on the same clip instead of regenerating from scratch (works across full 15s)
  • Speech — more natural dialogue, clearer lip-sync, distinct multi-speaker conditioning
  • Camera — continuous FPV / orbit / zoom moves with fewer accidental jump cuts

Partners already wiring it in production include Artlist, FLORA, invideo, Krea, Picsart, Photoroom, QuickFrame AI, and Runway (customer quotes on the Cloud blog).

Where to run it today

  1. Gemini API / Google AI Studio — prototype and ship against the GA model
  2. Google Cloud — console + API with enterprise SLAs / Provisioned Throughput
  3. Flow — scene direction and multi-turn edits
  4. Google Vids — 15-second clips at no additional cost with a Google account or Workspace
  5. Gemini app — 15-second video rolling out to Google AI Ultra subscribers
  6. Google Ads — scale campaign creatives

ArtRealmAI Gen does not host Gemini Omni — this stays on Google’s API and apps.

Caveats

  • Closed weights — no local Comfy twin
  • Audio reference inputs remain allowlist-gated on Cloud
  • Google still flags sports action, rapid script changes, and complex physics as work-in-progress
  • Cost is per-second of video; drafting low-res first is the intended cost control
  • Prior ArtRealm coverage of Omni 1.1 / Vids free HD is upstream context — this piece is the 1.2 Flash GA drop

Original Source

https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemini-omni-1-2-flash/

Conclusion

Fifteen seconds, cheaper seconds, and conversational edits — Omni 1.2 Flash is the GA video lane Google wants brands and apps to pin in production today. If you already draft in AI Studio, Flow, or Vids, this is the upgrade to flip on; if you live in open Comfy graphs, treat it as a strong closed reference surface, not a weights drop.

—Aurelia ♡