Google introduced Gemini Omni at I/O 2026, and Gemini Omni Flash is the fast, low-cost model in the family. It creates video from almost any input, then lets you keep editing the clip just by chatting.
The model has moved quickly since then. Gemini Omni 1.1 Flash added 4K output, start and end frames, and reference videos.
Gemini Omni 1.2 Flash raised the clip length to 15 seconds, made upscaling its own tool, and cut the price by about 30%.
If you have used Nano Banana for images, this is the video counterpart: fast, flexible about what you feed it, and grounded in Gemini's understanding of the real world. Google even describes it as "Nano Banana for video."
Below is a plain explanation of what Gemini Omni Flash is, what changed in 1.2 Flash, what you can make with it, how it fits Google's lineup, and where to try it.
Key takeaways
- Gemini Omni Flash is the fast, low-cost video model in Google's Gemini Omni family, and Gemini Omni 1.2 Flash is the current version.
- It takes text, images, and video, and makes clips up to 15 seconds long with native audio, from 360p up to 4K.
- Its standout feature is conversational editing: refine a video up to 15 seconds long by describing the change.
- Upscaling is a standalone tool in 1.2 Flash, so any clip up to 15 seconds long can go to 1080p or 4K.
- It costs about 30% less than 1.1 Flash, and you can use it on OpenArt alongside Veo 3.1, Seedance 2.5, and Kling 4.0.
What is Gemini Omni Flash?
Gemini Omni Flash is Google's fast AI video model, the Flash line of the Gemini Omni family. Gemini Omni 1.2 Flash (model ID gemini-omni-1.2-flash) is the current version.
It turns text, images, and short video clips into finished footage up to 15 seconds long, complete with synchronized sound.
What sets it apart is that it handles text, images, audio, and video together, so the picture, the sound, and your edits all line up. You shape a clip step by step through natural conversation, and the parts you did not mention stay the same.
Gemini Omni 1.2 Flash vs 1.1 Flash
Gemini Omni 1.2 Flash keeps everything 1.1 Flash could do and changes four things.
Longer clips: up to 15 seconds
Text to video and image to video clips run up to 15 seconds long, up from 10 on 1.1 Flash. Clips made with a start and end frame, or with reference images and videos, stay at 10 seconds.
Longer edits: videos up to 15 seconds
Conversational editing takes videos up to 15 seconds long, up from 10. You upload the clip, describe the change, and the rest of the scene stays as it was.
Standalone upscaling to 1080p or 4K
Upscaling is its own tool in 1.2 Flash. Any clip up to 15 seconds long, from Omni Flash or anywhere else, can be upscaled to 1080p or 4K without being made again.
About 30% lower cost
Every resolution, from 360p to 4K, costs about 30% less than it did with 1.1 Flash. That makes the draft-then-upscale habit cheap: make the first versions at 360p or 720p, then upscale only the keeper.
Google also says 1.2 Flash follows image and video references more closely, so products, characters, and brand details stay consistent from shot to shot. Voice lines sound more natural too, with clearer lip-sync and distinct voices when more than one person speaks.
What you can actually do with it
Gemini Omni Flash covers four core AI video jobs: text to video, image to video, conversational editing, and upscaling, with native audio on every clip it makes. That range means a single model handles most of a short-form workflow, from first idea to final 4K file.
Here is where it helps real teams:
Generate from a prompt:
Describe a scene and get a clip up to 15 seconds long with audio, at 360p, 720p, 1080p, or 4K.
Animate photos:
Bring a still image or product shot to life with realistic motion, or set a start frame and an end frame and let the model make the move between them.
Edit by chatting:
Swap a character, change the lighting, or restyle a clip up to 15 seconds long by describing it, while the rest stays intact.
Hold consistency with references:
Point to up to 10 reference images and short reference videos for a character, product, style, or composition.
Upscale any clip:
Take a draft made at 360p or 720p, or a video from anywhere else, up to 1080p or 4K without making it again.
Get readable on-screen text:
Put the words in your prompt and Omni Flash places them cleanly in the video.
Because it is grounded in Gemini's knowledge of physics, history, and culture, prompts tend to land accurately even when they are short and simple.
Gemini Omni 1.2 Flash at a glance
| Spec | Gemini Omni 1.2 Flash |
|---|---|
| Clip length | 3 to 15 seconds for text to video and image to video; up to 10 seconds with a start and end frame or with reference images and videos |
| Editing input | Videos up to 15 seconds long |
| Resolutions | 360p, 720p, 1080p, and 4K |
| Standalone upscaling | Any clip up to 15 seconds long, to 1080p or 4K |
| Aspect ratios | 16:9 and 9:16 |
| References | Up to 10 images plus up to 3 short video clips |
| Audio | Native sound, including voice lines, sound effects, and music |
| Price | About 30% less than Gemini Omni 1.1 Flash at every resolution |
Where Gemini Omni Flash sits in Google's AI models lineup
Omni Flash is the speed-focused, cost-efficient tier of the Omni family, built for high-volume work where you make many versions. It is the video model in the Gemini app, and it also powers video in Google Flow and Google Vids.
The "Flash" name signals the same thing it does across Gemini: fast and affordable, tuned for quick turnaround. With 15-second clips and a 4K upscaler in 1.2 Flash, that trade covers most short-form video.
The easiest way to try Gemini Omni Flash: OpenArt
The simplest way to try Gemini Omni 1.2 Flash is on OpenArt, right in your browser, with one account that covers every model. Pick it in the model menu of OpenArt's AI video generator, write a prompt, and generate.
The bigger advantage is that OpenArt runs many video models under one plan. Alongside Gemini Omni 1.2 Flash, you can switch between Veo 3.1, Seedance 2.5, Kling 4.0, and more, then finish clips with Extend Video or the AI video upscaler.
Generate with Omni Flash when you want speed, 15-second clips, and chat-based edits, and jump to another model when a project needs something different. For a multi-scene video up to 5 minutes long, Director mode plans the scenes and builds the whole video.
Frequently asked questions
What is Gemini Omni Flash built on?
It is the Flash line of Google's Gemini Omni family, a video generation and editing model that takes text, images, and video as input. The current version is Gemini Omni 1.2 Flash (gemini-omni-1.2-flash).
How long are the videos?
Up to 15 seconds per generation with Gemini Omni 1.2 Flash, with native audio. Clips made with a start and end frame, or with reference images and videos, run up to 10 seconds, and on OpenArt you can add length with Extend Video.
What is the difference between Gemini Omni 1.2 Flash and 1.1 Flash?
1.2 Flash makes and edits clips up to 15 seconds long instead of 10, upscales any clip to 1080p or 4K as a standalone tool, and costs about 30% less at every resolution. Google also says it follows references more closely and makes more natural voice lines.
Is it the same as Gemini 2.0 Flash?
No. Gemini 2.0 Flash was a language model. Omni Flash is a dedicated 2026 video model in the Gemini Omni family by Google.
Can I edit a clip without regenerating it?
Yes. Describe the change in chat, and Omni Flash applies it while keeping the rest of the scene, motion, and camera position the same. With 1.2 Flash, the clip you edit can be up to 15 seconds long.
Can I upscale a video I already have?
Yes. In OpenArt's AI video upscaler, pick Gemini Omni 1.2 Flash, upload a clip up to 15 seconds long, and choose 1080p or 4K.
Where can I use it?
Through Google's Gemini app, Google Vids, and the Gemini API, or on OpenArt alongside other top video models on one platform.