VideoGoogle

Gemini Omni

Gemini Omni is Google's reference driven video model, taking up to 4 images in and generating picture with native audio. Reach for it when what is on screen has to match assets you already own.

Duration4, 6, 8, 10s

Resolutions720p, 1080p

Aspect ratios16:9, 9:16

AudioNative

Reference inputsUp to 4

Make with Gemini Omni →View pricing

Added July 2026. Included on every plan.

Why this one

What makes Gemini Omni different

Built around referencesIt takes up to 4 images as its primary input, rather than treating them as an extra.
Native audioSound is generated with the picture, so it lines up with the action.
No frame inputs to fightThere is no first or last frame here: identity comes from references, which keeps the mental model simple.
A four step duration ladderIncluding the longest option among the Google tiers.
A seed you controlPin it to reproduce a take when a client asks for one small change.
Flat pricing across resolutionsBoth resolution tiers cost the same, so there is no reason to draft low.

Use cases

Best for

Product in scene
Brand consistency
Character continuity
UGC style ads
Asset driven campaigns
Explainers with real products

Technique

Prompt tips

Lead with what the references areTell it which image is the product, which is the person, which is the place. It uses that.
Give clean, uncluttered referencesUp to 4, and quality matters more than quantity. Two sharp stills beat four busy ones.
Write the audioNative sound only helps if you say what should be heard.
Pin the seed for revisionsSmall changes to a good take are far easier with a fixed seed.
Pick the higher resolutionThe two tiers cost the same, so there is no saving in drafting low on this model.

Mechanism

How Gemini Omni works

You hand it up to 4 reference images and a prompt, and it composes a scene that uses those references as its subjects, generating matching audio at the same time. There are no frame anchors, so timing is directed in words while identity comes from the pictures, which is a cleaner split than most models offer.

Getting started

How to use it on Dolly

Pick Gemini OmniChoose it in the video studio, or just ask the agent for it by name.
Write the promptDescribe what you want, and add your reference images or frames where they help.
GenerateThe exact credit cost sits on the button before you press it, and a failed run refunds automatically.

Pricing

What it costs

Every generation shows its exact credit cost on the button before you run it.

720p · 8s: 84 credits
1080p · 8s: 84 credits

One run is about 1/15 of a Starter month.

Failed generations refund automatically. See plans

Capabilities

Specs

Resolutions: 720p, 1080p
Aspect ratios: 16:9, 9:16
Durations: 4, 6, 8, 10s
Reference images: Up to 4
Audio: Native
Settings: Seed
Max per run: 2

Questions

Gemini Omni FAQ

How many reference images can I give it?

Up to 4 per run, and quality matters more than quantity: two clean stills usually beat four busy ones.

Can I set a first or last frame?

No. This model uses references rather than frame anchors, so identity comes from the pictures and timing comes from the prompt.

Does it generate sound?

Yes, natively with the picture, which is why the audio lines up with what happens on screen.

Which resolution should I choose?

The higher one. Both tiers are priced the same, so there is no reason to draft lower.

Can I reproduce a take?

Yes, fix the seed and keep the prompt and references identical.

Ready to make with Gemini Omni?

Every model is included on every plan. The price sits on the button before you run it.

Start making →