Which model should I use for a product video?

Start with Veo 3.1. Upload a product image as reference, write a prompt describing the scene and camera movement, and generate at Standard quality first. Move to Pro or 4K after the direction is confirmed.

When should I use Veo 3.1 instead of Veo 3.1?

Use O3 when the output must closely match a reference image — for brand consistency, character identity, or specific composition. If you want more creative freedom and general video quality, use Veo 3.1.

What is Motion Control for?

Motion Control transfers movement from a reference video to your subject. Use it for dance, gesture, pose, or camera movement when you have a specific motion you want to replicate. It is not a general video generator.

Should I generate images before video?

Yes, if you need consistent reference frames. Use O3 Image to create the visual first, then use that image as a reference for video generation. This gives you more control over the final result.

When should I use 4K mode?

Only after the direction is locked. 4K costs more credits and is meant for final delivery. Draft at Standard or Pro first, confirm the shot works, then re-generate at 4K for the final output.

How do I avoid wasting credits?

Three rules: choose the right model for your task, start at Standard quality for iteration, and preview the credit cost before every generation. Most wasted credits come from using the wrong model or generating at 4K too early.

Can I preview the credit cost before generating?

Yes. The credit counter updates as you change settings — duration, resolution, quality mode, and audio. You always see the exact cost before submitting. Failed system generations are not charged.

What inputs does each model need?

Veo 3.1: text prompt, optional image. Veo 3.1: text prompt + reference image. Motion Control: subject image + motion reference video. O3 Image: text prompt, optional reference. O3 Bild bearbeiten: source image + edit prompt.

How do @element tags work?

Elements are saved visual references — characters, products, logos, or objects. Upload 2-4 images to create an element, then insert it into prompts with @element_name. This helps maintain consistency across multiple generations.

What is the maximum video duration?

Video duration depends on the model and quality mode. Most workflows support 5s or 10s clips. Multi-shot sequences can combine up to 6 shots within a 3-15 second total duration.

Gemini Omni

Gemini Omni Video Generator: Choose the Right Model Before You Generate

The model you pick matters more than the prompt you write. Start by choosing the workflow that matches your output — video draft, reference-guided clip, motion transfer, image asset, or 4K final — then generate with the right Gemini Omni model.

Start a Veo 3.1 Draft

Preise ansehen Plans

Preview your credit cost before every generation. One shared credit pool covers all models — video, image, motion, and 4K.

Need video?

Use Veo 3.1 for text-to-video or image-to-video drafts with strong motion quality

Need reference control?

Use Veo 3.1 when the output must match a reference image for style, identity, or composition

Need motion transfer?

Use Motion Control to apply movement from a reference video to your subject

Preview credits before every generation

No surprises. See the exact credit cost before you submit — then decide whether to iterate cheaply or commit to a higher-quality render.

Cost shown before you generate

The credit counter updates as you change duration, resolution, quality mode, and audio settings. You always know the cost before committing.

One credit pool for all models

Video, image, motion control, and 4K all draw from the same balance. No separate wallets or confusing tier splits.

Failed generations are not charged

If the system fails to produce output, credits are not deducted. You only pay for completed results.

Start cheap, upgrade when ready

Use Standard or Draft modes for iteration at lower cost. Move to Pro or 4K only when the direction is locked and the output needs to ship.

Which mode should I choose?

Pick the Gemini Omni model that matches your task

Most bad AI video drafts fail before generation starts — because the user picked the wrong model for the job. Use this guide to choose the right workflow before you write a single prompt.

Video generation (Veo 3.1 / O3)

For text-to-video, image-to-video, multi-shot sequences, and reference-guided clips. Choose Veo 3.1 for general video or O3 when reference fidelity matters.

Motion & Image workflows

For movement transfer from reference video, or creating still image assets before video generation.

Complete Workflow

Define the output

Video clip, reference-guided video, motion transfer, image asset, or image edit.

Choose the model

Veo 3.1, O3, Motion Control 3.0, Motion Control 2.6, O3 Image, or O3 Bild bearbeiten.

Add inputs

Prompt, reference images, @element tags, motion reference video — whatever the model needs.

Preview credits and generate

Check the cost, choose quality level, then generate. Iterate cheaply before committing to Pro or 4K.

For general video

Veo 3.1 handles most text-to-video and image-to-video tasks. Strong motion, good prompt following, Standard/Pro/4K quality tiers.

Veo 3.1

For reference fidelity

Veo 3.1 locks style, identity, and composition from a reference image. Use when brand consistency or character identity matters.

Veo 3.1

Model decision guide

Which Gemini Omni model should I choose?

Each model solves a different problem. Choosing wrong wastes credits and produces off-target results. Use this guide.

Veo 3.1 — General video generation

Your default choice for most video tasks. Handles text-to-video, image-to-video, and multi-shot sequences with strong motion quality and good prompt adherence.

Best for: product clips, ad variations, social content, creative exploration

Inputs: text prompt, optional reference image, optional @elements

Quality: Standard, Pro, and 4K modes available

Duration: 5s or 10s with 16:9, 9:16, and 1:1 ratios

Cinematic Scene Generation

An AI-generated cinematic scene demonstrating Veo 3.1 text-to-video output.

Dynamic Motion Showcase

A dynamic motion clip showcasing AI video generation with fluid movement.

Veo 3.1 — Reference-guided video

Use when the output must closely match a reference image in style, composition, or character identity. Stronger reference fidelity than Veo 3.1.

Best for: brand consistency, character identity, style-locked sequences

Inputs: text prompt + reference image (required for best results)

Quality: Standard and Pro modes

When to use over 3.0: when reference matching matters more than creative freedom

Motion Control — Movement transfer

Not a general video model. Use specifically when you have a still subject and want it to follow movement from a reference video.

Best for: dance, gesture, pose, camera movement, product animation

Inputs: subject image + motion reference video (both required)

Models: Veo 3.1 or Veo 3.1

When NOT to use: open-ended video generation — use Veo 3.1 instead

O3 Image — Still image generation

Create reference frames, product visuals, thumbnails, or style frames before moving to video generation.

Best for: reference images, product concepts, thumbnails, style exploration

Inputs: text prompt, optional reference image

Output: 1K/2K or 4K still images

Workflow tip: generate images first, then use them as video references

O3 Bild bearbeiten — Modify existing images

When an image is close but needs prompt-guided changes before it becomes a reference or final asset.

Best for: background swap, object changes, style adjustments, cleanup

Inputs: source image + edit prompt (both required)

Output: edited image at source resolution

Workflow tip: edit first, then use the result as a video reference

4K Mode — Delivery-grade output

Not a separate model — a quality tier available on supported workflows. Use only after the direction is locked.

Best for: final delivery, broadcast, client presentations, portfolio

When to use: after Standard/Pro drafts confirm the direction works

Cost: higher credits than Standard/Pro — do not use for iteration

Rule: draft first at lower cost, 4K only when the shot is approved

Workflow

From task to generated draft in 4 steps

Start with what you need, not which button to press. The right model choice saves credits and produces better results on the first try.

Name the output you need

Video clip? Reference-guided video? Motion transfer? Image asset? Image edit? Start with the result.

Task

Pick the matching model

Veo 3.1 for general video. O3 for reference control. Motion Control for movement. O3 Image for stills.

Model

Add inputs and preview credits

Write the prompt, upload references, check the credit cost. Adjust quality settings before generating.

Setup

Generate, compare, iterate

Start with Standard quality for fast iteration. Move to Pro or 4K only when the direction is confirmed.

Generate

Choose by task, not by tool name

The fastest path to a good result is choosing the right model before you write the prompt. Here is when to use each.

I need a video clip from a prompt or image

Use Veo 3.1. It handles most video generation tasks with strong motion quality. Add a reference image if you want visual guidance, or use text-only for creative exploration.

→ Veo 3.1

I need the output to match a specific reference closely

Use Veo 3.1. It prioritizes reference fidelity over creative freedom. Best when brand identity, character consistency, or composition must be preserved.

→ Veo 3.1

I need to transfer movement from one video to my subject

Use Motion Control. Upload your subject image and a motion reference video. The model applies that movement to your subject.

→ Motion Control

I need an image before I start video work

Use O3 Image to generate reference frames, product visuals, or style explorations. Then use those images as references for video generation.

→ O3 Image

Answers

Gemini Omni Generator FAQ

Answers for choosing the right Gemini Omni model and workflow.

Choose your workflow and start generating

Pick the model that matches your task. Preview credits before every job. Iterate cheaply, then upgrade to Pro or 4K when the direction is locked.

Start a Veo 3.1 Draft

Create Reference-Guided Video

Transfer Motion