Several images, one video
Add up to 7 reference images: a product, an outfit, a character, a setting. Google built Gemini Omni to combine several sources into one coherent clip; say in the prompt what each image is for.
Video model · Google
Gemini Omni — Short clips up to 4K, guided by your images.
Gemini Omni is Google’s video model family, announced in May 2026; its first model, Gemini Omni Flash, brings text and images together in one video, with sound. On Deepnia it is the default video model: 4, 6, 8 or 10 s clips in 720p, 1080p or 4K, 9:16 or 16:9, with speech, music and sound effects, and up to 7 reference images to guide a product, an outfit or a setting.
- 4, 6, 8 or 10 s
- Up to 4K
- Sound included
- 7 reference images
- Default video model
The video tool opens with Gemini Omni already selected. The exact cost shows before every creation; you sign up when you generate.
- A text prompt
- Up to 7 images (optional)
- A 4–10 s clip, up to 4K
At a glance
- Maker
- Length
- 4, 6, 8 or 10 s (6 s by default)
- Resolution
- 720p, 1080p, 4K (720p by default)
- Aspect ratios
- 9:16, 16:9
- References
- Up to 7 images
- Sound
- Included: speech, music, sound effects
What Gemini Omni does best
From 720p to 4K
Pick 720p, 1080p or 4K to suit the use, 720p by default. Google states that 1080p and 4K are produced by upscaling the generated video.
Readable text on screen
Google says Gemini Omni renders requested text correctly, such as a label or word-by-word captions, in sync with the action. Keep that text short.
A believable physical world
Google highlights the model’s grasp of physical forces (gravity, momentum, fluids) and of real-world knowledge, from history to science: useful for explainer videos.
Several shots in one clip
According to Google, Gemini Omni Flash chains several shots by default. You can direct the cuts and the timing in the prompt, or ask for one continuous shot.
Videos that talk, in French too
Every video comes with its own sound: voices, music and sound effects. Write your prompt in English or French and put the lines in quotation marks to make your characters speak.
How to use Gemini Omni on Deepnia
Open the video tool
The “Try Gemini Omni” button opens Deepnia’s video tool with the model already selected; it is also the default video model.
Add your images if you need them
Attach up to 7 photos from your device or your Deepnia library (10 MB per image). Without images, the video starts from text alone.
Describe the scene
State what each image is for, the camera movement and, if needed, the timing (“[0-3s] …”).
Set it up and generate
Choose the length, resolution and aspect ratio, check the cost shown, then create.
Gemini Omni on Deepnia: what you can set
- Inputs
- Text only, or text plus 1 to 7 reference images
- Length
- 4, 6, 8 or 10 s (6 s by default)
- Resolution
- 720p, 1080p, 4K (720p by default)
- Aspect ratios
- 9:16, 16:9
- References
- Up to 7 images of up to 10 MB, from your device or Deepnia library
- Sound
- Generated with the video: speech, music and sound effects
- Prompt
- Required, in English or French, up to 5,000 characters
Prompt ideas to try
Copy a prompt, open the tool and adapt it to your project.
Product teaser for a drinks brand
- 9:16
- 8 s
- 1080p
- 1 product photo
One continuous shot, no cuts. Close-up of the bottle from the attached image on a wooden counter, condensation droplets on the glass, golden late-afternoon light. The camera slowly pulls back to reveal a busy café in the background. Realistic commercial look, no added text.
Animated lookbook for a fashion brand
- 9:16
- 10 s
- 1080p
- 3–5 outfit photos
Use the attached images as outfit references, not as first frames. Three models walk one after another across a terrace at sunset; each wears one of the reference outfits, with identical patterns and colors. Shoulder-height camera following the walk, fashion magazine style.
Concert announcement with short text
- 16:9
- 6 s
- 1080p
- Text only
Open-air concert at night, facing the sea. The crowd raises their arms under purple and orange lights. In the last two seconds, large white text appears in the center of the frame: "SATURDAY 8 PM". Only one text on screen, clearly readable.
Classroom clip on the water cycle
- 16:9
- 10 s
- 720p
- Text only
Educational animation in colorful 3D illustration: the water cycle above a coastal village. [0-3s] The sun heats the lagoon and light vapor rises. [3-7s] Clouds form above the palm trees. [7-10s] Rain falls on the roofs and flows back into the lagoon. Three short labels appear at the right moment: "evaporation", "condensation", "rain".
From sketch to video for an illustrator
- 16:9
- 8 s
- 720p
- 1 sketch
Turn this sketch into a realistic sequence, using the drawing only as a guide for the motion and never showing the drawing: a bike courier weaves between cars on a busy avenue, golden evening light, low tracking camera. One continuous shot.
Evening mood at a grill restaurant in Kigali
- 16:9
- 8 s
- 4K
- Text only
One continuous shot, no cuts: a slow tracking shot between the tables of a rooftop grill restaurant in Kigali, in the evening. Flames on the grill, golden grilled fish, guests chatting under string lights. Warm light, cinematic look, no on-screen text.
Tips for better results
- Gemini Omni Flash chains several shots by default: write “one continuous shot, no cuts” when you want a single take.
- Say what each image is for (“as an outfit reference”, “not as the first frame”): your prompt is what guides the model.
- To make a character speak, put the line in quotation marks and describe the tone (“in a calm voice”); describe the music or ambience you want as well.
- Set the pacing with time markers (“[0-3s] … [3-7s] …”): Google documents this syntax in its guide.
- For on-screen text, put it in quotation marks, keep it short and describe any background text as well.
- Write what you don’t want to see straight into the prompt, for example “no on-screen text”.
- Validate the idea at 720p and 4 s, then switch to 4K for the final cut.
Good to know
- To tell a longer story, create several 10-second clips and join them in your editor.
- Pick 9:16 for Reels, TikTok and Shorts, and 16:9 for YouTube and the web.
- To keep a character identical from shot to shot, attach their photo as a reference, stay in one scene and avoid fast pans.
- Use images you have the rights to: Google may refuse photos of some well-known people and, in some regions, of minors.
Gemini Omni or another model?
Deepnia’s other video models, to pick the right one for your project.
| Model | Length | Resolution | Aspect ratios | Best for |
|---|---|---|---|---|
| Gemini OmniThis model | 4, 6, 8 or 10 s | 720p–4K | 9:16 · 16:9 | Short clips up to 4K, guided by your images. |
| Veo | 8 s | Set by the mode | 9:16 · 16:9 | 8 s clips with sound, in three quality levels. |
| Seedance 2.5 | 4–30 s | 480p–720p | 16:9 · 9:16 · 1:1 · 3:4 · 4:3 · 21:9 | Videos up to 30 s with sound and up to 16 image and video references. |
| Kling 3.0 | 3–15 s | Set by the mode | 9:16 · 1:1 · 16:9 | A 5-shot editor, persistent characters and optional sound. |
| HappyHorse | 3–15 s | 720p–1080p | 9:16 · 1:1 · 16:9 · 4:3 · 3:4 | Editing an existing 3–60 s video. |
| Grok Video | 6–30 s | 480p–720p | 9:16 · 1:1 · 16:9 · 2:3 · 3:2 | 6–30 s videos, also in 1:1, 2:3 or 3:2. |
Frequently asked questions
What is Gemini Omni?
A Google video model family, announced on May 19, 2026 at Google I/O. Its first model, Gemini Omni Flash, creates video from text, images and video. On Deepnia it is the default video model, used with text and up to 7 images.
How long can a Gemini Omni video be on Deepnia?
4, 6, 8 or 10 seconds, 6 seconds by default. For a longer video, create several clips and join them in your editor.
Can I create videos in 4K?
Yes: Deepnia offers 3 resolutions (720p, 1080p, 4K), 720p by default. Google states that 1080p and 4K are produced by upscaling the generated video.
Can I start from my own photos?
Yes, up to 7 images from your device or your Deepnia library, 10 MB each. The model decides how to use each image based on your prompt, so spell it out, for example “as an outfit reference”.
Can Gemini Omni make characters speak French?
Yes. Write your prompt in English or French and put the lines in quotation marks: Gemini Omni makes your characters speak, with music and sound effects.
Do Gemini Omni videos have sound?
Yes. Every video is generated with its own sound: speech, music and sound effects, with nothing to switch on. Describe the voices and the ambience you want in the prompt.
How do I keep the same product or character across several videos?
Attach the same reference images to each creation and describe the product or character with the same words. One scene per clip gives the most consistent results.
Gemini Omni or Veo: which one should I use?
Gemini Omni to choose the length and the resolution up to 4K, or to combine up to 7 images. Veo for 8-second clips in three quality levels.
Ready to try Gemini Omni?
The video tool opens with Gemini Omni already selected. The exact cost shows before every creation; you sign up when you generate.