Google made its Gemini Omni 1.2 Flash video model generally available on Thursday, November 5: clips now run up to 15 seconds, editing happens in conversation instead of starting over, and 720p pricing drops 30% to about $0.07 per second — a combination aimed squarely at production video work, not demos.
November 5, 2026
What happened
Google announced the general availability of Gemini Omni 1.2 Flash, its video-generation and editing model, on the paid Gemini API and Google Cloud, with prototyping in Google AI Studio. The model generates synchronized video and native audio from text, image, and video inputs. Native clip length rises from 10 seconds in Omni 1.1 to 15 seconds, settable in whole-second steps from 3 to 15, with 720p and 16:9 as defaults, 9:16 portrait supported, and the option to draft at 360p or 720p and upscale selected results to 1080p or 4K.
Editing without regenerating
The headline workflow change is multi-turn conversational editing: users can replace a product or a background, change lighting or weather, or switch aspect ratio while leaving everything unmentioned intact, without re-uploading the prior video. Each request can use up to six reference images, including up to three images per character for up to two characters. Google DeepMind’s side-by-side evaluations give Omni 1.2 a 62% visual-consistency win rate against 38% for Omni 1.1 in text-to-video testing, and a 63% multi-reference identity-consistency win rate. Video extension is separate: it appends 3 to 10 seconds per turn, up to 30 seconds total.
Pricing and migration
On the paid Gemini API, video output is priced at $12 per million video-output tokens, which works out to roughly $0.70 for a 10-second 720p clip and $1.04 for a 15-second clip; text, image, and video input costs $1.50 per million tokens, and the model is not available on the free tier. Omni 1.1 Flash stays available for a four-week transition — on December 3, 2026, API requests to 1.1 will automatically route to 1.2 at the lower rates, with no code changes required.
Where it shows up
Beyond the API, the rollout covers Flow for scene direction and multi-turn edits, longer clips in Google Vids at no additional cost with a Google account or Workspace subscription, a Gemini app rollout for Google AI Ultra subscribers, and campaign creation in Google Ads. Every video generated through the API carries an invisible SynthID watermark. Photoroom is already building on it in production: its Video Ad Maker and Product Animator run on Gemini Omni Flash, and chief executive Matt Rouif said the longer generation limit was a top user request. Google is candid about the remaining limits — fast sports action, rapid script changes, and complex physics are still in development; the API fully supports English but has not evaluated other languages; and voice editing and uploaded audio inputs are not supported in the API.
Why it matters
AI video’s economics change when a fix costs an edit instead of a fresh generation. Fifteen seconds is long enough for a complete social ad or product explainer in one pass, and a 30% price cut plus a dated, automatic migration path tells production teams Google wants this model treated as infrastructure, not a novelty. The consistency numbers — if they hold up outside Google’s own testing — are the real story: keeping the same face, garment, and packaging readable across cuts is what has kept AI video out of serious brand work.
Frequently asked questions
What is the longest video Gemini Omni 1.2 Flash can generate?
Up to 15 seconds natively, in whole-second steps from 3 to 15 seconds. A separate extension feature can append 3 to 10 seconds per turn, up to 30 seconds total.
What does Gemini Omni 1.2 Flash cost?
About $0.07 per second at 720p — roughly $0.70 for a 10-second clip and $1.04 for 15 seconds — on the paid Gemini API. It is not available on the free tier.
What happens to Gemini Omni 1.1 Flash?
It remains available during a four-week transition. On December 3, 2026, API requests to 1.1 automatically route to 1.2 at the lower 1.2 rates, with no code changes needed.
Sources: Unite.AI, Google, Google Cloud