Gemini Omni 1.2 Flash is generally available
TL;DR
On November 5, 2026 Google released Gemini Omni 1.2 Flash as gemini-omni-1.2-flash. Video generation goes up to 15 seconds, video output is priced 30% lower at $12 per million tokens, and paid-tier rate limits doubled. Requests to Omni 1.1 Flash route to 1.2 on December 3.
- Status
- Live
- Released
- Plans
- Gemini API paid tier; Gemini app for Google AI subscribers
What changed
The Gemini API release notes for November 5, 2026 mark gemini-omni-1.2-flash generally available. Google lists:
- Video generation up to 15 seconds.
- Conversational editing: 3 to 10 seconds for first-and-last-frame interpolation, and 3 to 10 seconds of video added per turn, up to 30 seconds total.
- More consistent characters and subjects across up to 6 reference images.
- Video output pricing 30% lower, at $12 per million tokens.
- Paid-tier rate limits doubled.
On December 3, 2026, requests to gemini-omni-1.1-flash start routing automatically to gemini-omni-1.2-flash and are billed at 1.2 rates.
@GoogleAI's November 5 post describes the same model as more cost-effective, with a 15-second generation and editing limit, scenes that hold together longer, better character speech, fewer jump cuts, and closer adherence to reference assets. @GeminiApp said the model is rolling out to all Google AI subscribers, with conversational edits and sharper visual stability in the Gemini app. It also said Google AI Ultra subscribers can create and edit videos up to 15 seconds.
The same day's API notes cover a separate change to Office files and function responses. That is its own feed entry.
Why it matters
If you generate video on Omni 1.1, you have until December 3 before those calls silently become 1.2 and pick up 1.2 prices. The unit Google published is $12 per million output tokens, 30% below the previous video output price. It did not publish a side-by-side rate-limit table, only that paid-tier limits doubled.
Fifteen seconds is the generation cap in the API notes. Extending a clip is 3 to 10 seconds per turn and stops at 30 seconds total. In the Gemini app, the 15-second create-and-edit limit is what Google announced for Ultra subscribers, while the model itself is rolling out to all Google AI subscribers.
Reference images cap at 6. A pipeline that was stuffing more references into 1.1 should be checked against that limit.
Frequently asked questions
What is the model ID?
gemini-omni-1.2-flash.
When does Omni 1.1 start routing to 1.2?
December 3, 2026. Those requests are billed at Gemini Omni 1.2 Flash rates.
How long a video can I generate?
The API notes say up to 15 seconds for generation. Per-turn extension is 3 to 10 seconds, up to 30 seconds total. First-and-last-frame interpolation is 3 to 10 seconds.