- Published on
Gemini Omni 1.2 Flash: Edit Videos by Talking, 30% Cheaper
"What if you could edit a video by talking to it, the way you chat with a friend?"
Gemini Omni 1.2 Flash turns that idea into an API-level reality.
What Changed: 15 Seconds, Conversational Editing, 4K
Google released Gemini Omni 1.2 Flash on the Gemini API as a generally available model. Three changes stand out.
First, 15-second video generation. The previous version 1.1 was capped at 10 seconds. Version 1.2 raises that to 15 seconds. For social media clips and short-form ads, those extra five seconds matter.
Second, multi-turn conversational editing. This is the most significant shift. After generating a video, you can continue the conversation: "change the background in that scene," "brighten the lighting," "make the subtitle text white." The editing context carries forward through the conversation.
Previously, revising a generated video meant rewriting the prompt from scratch or returning to a video editor. Multi-turn editing removes that break in the workflow.
Third, 1080p and 4K upscaling. Generated clips can be upscaled to high resolution. A practical workflow emerges: draft quickly at lower resolution to test ideas, then upscale for the final output.
Prices Dropped 30%
Gemini Omni 1.2 Flash is priced 30% lower than version 1.1, at $0.07 per second for 720p video.
AI video generation still costs significantly more than text generation. Even so, a 30% reduction matters for the developers and small studios who want to test these models in production. For B2B use cases involving large volumes of short-form content — ads, educational clips, social posts — the economics shift meaningfully.
Integrated into Flow and Google Vids
Gemini Omni 1.2 Flash is not only available as a standalone API. It has been integrated into several Google products:
- Flow: Google''s AI video creation tool, now capable of longer projects using Omni 1.2 Flash.
- Google Vids: The workplace video app that lets teams produce internal content with AI-generated video built in.
- Gemini for Business: Available to enterprise Google customers.
The integration signals that AI video is moving from developer-only territory into tools used by everyday creators and office workers.
What This Means for Educational Content
In education and edtech, there is a large unmet need for short, targeted explainer videos. A teacher who wants a 15-second animation of a concept, or a training team that needs fifty short onboarding clips, cannot currently produce those affordably or quickly.
Conversational video editing changes the production equation. "Make a 15-second video explaining fractions, in an animation style, then change the character to look younger" — if that kind of iterative dialogue produces usable drafts, the barrier to producing educational video drops substantially.
Current limitations still apply: 15-second maximum length, regional restrictions in some markets, and no support for voice editing in uploaded videos. But the trajectory is clear.
A New Paradigm: Editing as Conversation
The biggest shift that text AI brought was removing the need to memorize commands — you just describe what you want. Gemini Omni 1.2 Flash brings that same shift to video editing.
Instead of dragging a timeline or adjusting layers, you watch the output and say "slow that part down" or "fade the background music." That is becoming a realistic option, not just a demo.
Tips
- Developers on the Gemini API: check the official docs for the current model ID before integrating (the docs page may still show
gemini-omni-1.1-flash). - Flow and Google Vids users: look for updated video generation options in the UI that now leverage Omni 1.2 Flash.
- Multi-turn editing is especially effective for creating A/B variants of short ad videos without regenerating from scratch.
Sources