Karyakarya

Guides

Gemini Omni 1.2 Flash Is Generally Available - What It Means for Your Video Workflow

Edison Chua5 min read
Guides

Google made Gemini Omni 1.2 Flash generally available on November 5, 2026, across Google AI Studio, Vertex AI and the Gemini API. The two headline changes are simple: clips can now run up to 15 seconds instead of 10, and 720p generation costs about 30% less than the previous version, at $0.07 per second of output. Google says 1.1 Flash stays available for now but is set to route to 1.2 by default on December 3, 2026.

Neither change is flashy on its own. But for anyone making short-form video for TikTok, Reels or YouTube Shorts with an AI model that leans on prompt accuracy and reference-guided generation - which is exactly what the Gemini Omni family is known for - both numbers are worth paying attention to.

Why 5 extra seconds matters more than it sounds

Most AI video models cap out well under 10 seconds per clip, which forces creators into a familiar workaround: generate several short clips and stitch them together in an editor to reach a usable length. Every stitch point is a place where motion, lighting or a subject's appearance can visibly shift between clips.

A model that can hold a single continuous shot for 15 seconds instead of 10 means fewer cut points for a given video length. A 30-second UGC-style ad that needed three generations at 10 seconds each now needs only two at 15 - one fewer seam to hide, one fewer chance for a reference image to drift off-model between clips.

Why the price cut matters for iteration, not just cost

A 30% drop in per-second video pricing changes behavior more than it changes a single invoice. When a test run is cheaper, people run more of them - trying a different camera direction, a different pacing note, a different lighting cue - before committing to the version they actually publish. That is the real value of a price cut on a generation model: it buys you more attempts at getting the prompt right, which is usually the actual bottleneck, not the render itself.

Where this leaves Gemini Omni in Karya today

Karya's Image and Video studios currently run Gemini Omni Flash 1.1 - the version Google is retiring from default routing in December, not the new 1.2 release. That is not a gap unique to Karya: any platform that routes generations through a specific model version needs to verify a new release actually works end-to-end (correct inputs, correct outputs, stable pricing) before switching a paying customer's generation onto it.

If you want to use Gemini Omni for prompt-accurate, reference-guided video right now, the 1.1 version in Karya's Video Studio still supports both text-to-video and image-to-video, with the credit cost shown before you generate. When 1.2 Flash clears Karya's onboarding check, expect the longer 15-second option and the lower per-second cost to show up as an update to the existing Gemini Omni entry, not a new model card.

The takeaway

This week's update is a reminder that the most useful AI model news for creators is not always a brand-new model - it is an existing one getting cheaper and less constrained. Whichever platform you generate on, when a video model you already use adds length or cuts price, it is worth revisiting shots you previously had to split across multiple clips.

Try it in Karya

Generate images and videos with frontier AI models from one canvas. Subscribe from $19/mo, or top up from $5 - 200 bonus credits with your first purchase, and top-up credits never expire.

Start free