Gemini Omni is the model family Google launched at I/O 2026 instead of a Veo 4. The first release, Gemini Omni Flash, takes text, images, video and voice in one prompt and returns a video you can then edit by chatting with it. It hit general availability in the API on 27 August 2026, and the preview endpoint switches off on 30 September 2026. Here is what that means for a short-form calendar.
Gemini Omni is Google's any-to-any generative model family, announced at Google I/O on 19 May 2026, starting with Gemini Omni Flash: it accepts text, images, video and voice references in a single prompt and generates 3 to 10 second videos that can be edited through conversation while keeping characters and physics consistent. It is in the Gemini app for Google AI Plus, Pro and Ultra subscribers, in Google Flow, free inside YouTube Shorts Remix and the YouTube Create app, and in the Gemini API as gemini-omni-1.1-flash (GA 27 August 2026) at roughly $0.10 per second of 720p video. The gemini-omni-flash-preview endpoint is deprecated on 30 September 2026.
For marketers the point is editing, not generation: you can keep a character and a scene and change the setting, style or camera by asking, which turns one good clip into a set of variants.
| Maker | Google (Google DeepMind) |
|---|---|
| Released | Announced 19 May 2026 at Google I/O, live in the Gemini app and Flow the same day; API preview (gemini-omni-flash-preview) 30 June 2026; GA as gemini-omni-1.1-flash on 27 August 2026 |
| Status | gemini-omni-flash-preview is deprecated on 30 September 2026; migrate to gemini-omni-1.1-flash. A higher-end Gemini Omni Pro has been teased with no release date |
| Inputs | Text, images, video and audio in one prompt. Audio input is limited to voice references at launch |
| Output | Video only for now; Google says image and audio outputs are coming in later releases |
| Max clip length | 3 to 10 seconds per generation; clips can be lengthened with the 1.1 video extension feature |
| Resolution | 720p default; the 1.1 API adds a resolution parameter for 360p, 720p, 1080p and 4K |
| Editing | Conversational multi-turn editing of generated or uploaded video (setting, style, camera angle), with character consistency across turns. Audio and speech editing of existing video is not supported yet |
| New in 1.1 | Video extension (continue a clip from its end) and interpolation (generate a transition between two images) |
| Native audio | Not verified: Google's launch materials do not detail synchronized soundtrack generation, and audio output is listed as coming later |
| Pricing (API) | $1.50 per 1M input tokens, $17.50 per 1M video output tokens, billed at 5,792 tokens per second of 720p video, about $0.10 per second. No free API tier. Per-second prices at 1080p and 4K not verified |
| Pricing (apps) | Included in Google AI Plus, Pro and Ultra subscriptions; free to use inside YouTube Shorts Remix and the YouTube Create app |
| Watermark | Imperceptible SynthID watermark and an AI label on every clip, checkable in the Gemini app, Chrome and Google Search |
Gemini Omni is a new family rather than a Veo version: Veo 3.1 remains the current Veo and Google has not announced a Veo 4. Versus the June preview, gemini-omni-1.1-flash (27 August 2026) added video extension, two-image interpolation and a resolution parameter from 360p up to 4K, and became the stable production model.
The Gemini app and Google Flow for Google AI Plus, Pro and Ultra subscribers; YouTube Shorts Remix and the YouTube Create app at no cost; developers use gemini-omni-1.1-flash through the Gemini API (Interactions API).
The feature that matters for a brand account is conversational editing. Most video models give you a fresh roll of the dice every time you change the prompt; Omni keeps the character, product and physics from the last turn and changes only what you asked for. That is the workflow short-form actually needs: one approved clip, then five versions with a different setting, a different camera move and a different opening, for hook testing.
The free route through YouTube Shorts Remix makes it the cheapest way to try AI video natively on a short-form platform, but the clip cap of 10 seconds per generation and the uncertain soundtrack story mean it is a clip generator, not a finished-ad generator. Plan to cut, caption and add your own voiceover.
The 30 September 2026 preview deprecation is a small but real deadline: anything wired to gemini-omni-flash-preview stops working, and the migration is a model ID change plus a check on the new resolution parameter.
Generate anywhere, publish natively to TikTok, Reels, Shorts and LinkedIn, attribute signups and sales to every post.
Generate and edit hero clips in Gemini Omni, then upload them into Fastlane to schedule weeks ahead and publish natively to TikTok, Instagram Reels and YouTube Shorts, with signups and sales attributed to each post so you know which variant earned its place.
For the rest of the calendar, Fastlane generates from your website URL without depending on any one model: AI UGC videos with 1,000+ hyper-realistic characters, slideshows, hook + demo videos, memes and live trend remixes, reviewed in Blitz mode. Free plan, no credit card. Read our full Gemini Omni guide for the workflow in detail.
Google's any-to-any generative model family, announced at Google I/O on 19 May 2026, starting with Gemini Omni Flash, which turns text, images, video and voice references into short videos you can edit by conversation.
No. Veo 3.1 is still the current Veo and Google has not announced a Veo 4; Gemini Omni is a separate model family that Google launched for video instead.
In the API, gemini-omni-1.1-flash costs about $0.10 per second of 720p video ($17.50 per million video output tokens). In apps it is included in Google AI Plus, Pro and Ultra, and free inside YouTube Shorts Remix and the YouTube Create app.
3 to 10 seconds per generation, with a video extension feature in gemini-omni-1.1-flash for continuing a clip.
It is deprecated on 30 September 2026. Switch to gemini-omni-1.1-flash, the GA model released on 27 August 2026.
Specs verified against the sources above on September 28, 2026. Model access and pricing change often; check the maker's page before you build on any number here.
Join 50,000+ founders and marketers publishing the same video to TikTok, Instagram, YouTube and LinkedIn. Free to start.
