Video generation

Veo 4

by Google DeepMind

Google DeepMind's next video model.
Native 4K, longer clips, multi‑shot character consistency, and a cinematic camera language in a single prompt.

Veo 4

by Google DeepMind

Key features

Capability details on this page reflect the Veo 4 specifications expected from Google DeepMind's public roadmap and are refreshed as Google publishes additional model card detail.

Technical specifications

4K native

True 4K output, not upscaled from 1080p

15 to 30s

Single-pass clips, well past the 8-second Veo 3.1 base

Native, multi-speaker

Synced dialogue, ambient, effects, and music in one pass

ID-locked

Face, wardrobe, and voice held across every cut

Text, image, frames, refs

Multi-input control inherited from Veo 3.1

SynthID

Imperceptible AI-provenance watermark on every clip

Use cases

Multi-shot short films

Storyboard a 20 to 30 second sequence in one prompt and keep the same character, wardrobe, and lighting across every cut without rebuilding references.

Broadcast-grade commercials

Cinematic camera language and native 4K output fit hero product spots where the camera move and the texture both have to read on a big screen.

Film pre-visualization

Sketch a scene with real coverage, wide, medium, and close, before committing to a shoot day. Longer clips shorten the path from idea to previz.

Music video sequences

Tie multiple beats together with consistent characters and locations. Native multi-speaker audio means the cut sits on the track without heavy post.

Educational explainers

Walk through a process across several connected shots instead of one 8-second clip. Storyboard mode keeps the visual logic steady between beats.

Architecture walkthroughs

Longer continuous camera moves through a space with stable geometry and lighting, suited to real-estate, interior, and industrial design teams.

Prompt examples

Cinematic short

Lone astronaut walking across a dust-red plain at sunset, low handheld camera following from behind, wind catching the suit fabric, cinematic

Edit prompt

Brand commercial

Glass perfume bottle rotating on a dark marble surface, slow orbital camera, soft top light, refractive caustics, premium feel

Edit prompt

Music video beat

Singer walking through a rain-soaked Tokyo crosswalk at night, neon reflections on the wet ground, slow dolly forward, shallow depth of field

Edit prompt

Architectural walkthrough

Continuous dolly through a minimalist concrete house at golden hour, long shadows, dust motes in the light, calm pacing

Edit prompt

Educational explainer

Cross-section of a beating heart, accurate anatomy, slow rotation, soft studio lighting, focus pulls suggesting callouts

Edit prompt

Multi-shot story

Wide shot of a hiker reaching a ridge at dawn, then medium of the same hiker pulling out a map, then close on weathered hands tracing a trail

Edit prompt

Overview

Veo 4 is Google DeepMind's next-generation video model, the successor to Veo 3.1 in the Veo line. It generates true native 4K clips of 15 to 30 seconds in a single pass, with locked character consistency across cuts, storyboard sequencing for multi-shot scenes, cinematic camera language, and synchronized multi-speaker audio. Every output carries Google's SynthID watermark on by default.

What Veo 4 changes

Where Veo 3.1 focused on per-clip realism at 4K with accurate physics and native audio inside an 8-second window, Veo 4 moves the emphasis from a single shot to a connected scene. Longer clips, multi-shot characters that hold without manual references, a storyboard mode that treats beats as one sequence, and a richer cinematic camera vocabulary mean a finished scene can come out of one prompt instead of a dozen stitched generations.

Veo 4 on Morphic

On Morphic, Veo 4 sits in the video model picker alongside Veo 3.1, Gemini Omni, Seedance 2.0, Kling, and the rest of the video catalog. Switch the prompt bar to Video mode, pick Veo 4, attach any references, and run the same brief through every shipping model when you want to compare.

Simple pricing

Get started for free today, with the option to upgrade or cancel anytime.

Basic

$9/ month
billed as $0 per year

1100 monthly credits

1 user only

All models

Workflows

Standard

$24/ month
billed as $0 per year

3625 monthly credits

1 user only

All models

Workflows

Pro

$45/ month
billed as $0 per year

6350 shared monthly credits

1 user

+ up to 4 more at extra cost

All models

Workflows

Pro Max

$170/ month
billed as $0 per year

24650 shared monthly credits

1 user

+ up to 9 more at extra cost

All models

Workflows

Enterprise

For higher limits

Custom

pricing and billing terms

High-volume credits
Custom seat limits
All models
Workflows
Pricing Gradient

Free

For playing around

$0

forever free

Up to 20 credits
1 user only
Limited models
Workflows
Compare plan details

FAQs

What is Veo 4?
Veo 4 is Google DeepMind's fourth-generation video model and the successor to Veo 3.1. It generates true native 4K video clips of 15 to 30 seconds in a single pass, with locked character consistency across cuts, storyboard sequencing, cinematic camera control, and synchronized multi-speaker audio. Every output carries Google's SynthID watermark for AI provenance.
How long are Veo 4 videos?
Veo 4 generates clips of roughly 15 to 30 seconds in a single pass, a meaningful step up from the 8-second window of Veo 3.1. Longer scenes are built by chaining beats together in storyboard mode, which keeps characters, lighting, and continuity steady across cuts.
What resolution does Veo 4 output?
Veo 4 outputs true native 4K (3840×2160) rather than upscaling from a lower resolution. The native 4K path keeps fine detail in skin, fabric, foliage, and small motion intact, which is the kind of detail that usually breaks when video is upscaled after generation.
How do I use Veo 4 on Morphic?
Open Morphic, switch the prompt bar to Video mode, and pick Veo 4 from the model picker. Add a text prompt, attach reference images or frames if you want locked characters, and run the generation. For multi-shot scenes, write each beat as a separate prompt in storyboard mode and Veo 4 returns a connected sequence.
How does Veo 4 compare to Veo 3.1?
Veo 3.1 is Google DeepMind's photoreal flagship at 4K with native audio, 8-second clips, and physics-accurate motion. Veo 4 inherits those strengths and pushes three things forward: duration of 15 to 30 seconds per clip, native multi-shot character consistency without manual reference juggling, and a richer cinematic camera language.
How does Veo 4 compare to Seedance 2.0?
Seedance 2.0 accepts up to 12 mixed assets per generation with native music sync and flexible 4 to 15 second clip lengths. Veo 4 focuses on longer single clips, locked characters across cuts, multi-speaker dialogue audio, and cinematic camera control. Pick Seedance 2.0 for music-led multimodal shorts; pick Veo 4 for dialogue-driven multi-shot scenes.
How does Veo 4 compare to Kling 3.0?
Kling 3.0 leads on stylized motion and strong image-to-video performance at competitive resolutions. Veo 4 leads on true native 4K output, multi-speaker synchronized audio, locked multi-shot characters, and storyboard sequencing. For long-form narrative work with dialogue, Veo 4 is the stronger fit; for stylized motion experiments, Kling stays competitive.
Does Veo 4 support character consistency?
Yes. Veo 4 anchors characters against reference images and holds face, wardrobe, and voice steady across every shot in a sequence. A subject in shot one reads as the same person in shot ten without re-uploading the reference each turn.
Does Veo 4 generate audio?
Yes. Veo 4 generates synchronized native audio in the same pass as the video: multi-speaker dialogue, ambient sound, effects, and music. Audio flows across cuts inside a storyboard rather than resetting every clip, so a dialogue scene reads as one continuous take.
Does Veo 4 include a watermark?
Yes. Every Veo 4 clip carries Google's imperceptible SynthID watermark. The watermark is invisible to viewers, on by default, and survives common edits like re-encoding and resizing for AI provenance.

Other models

Explore the rest of the Morphic model catalog.

See all models →

ChatGPT Images 2.5

OpenAI

OpenAI's image model, released September 2026. Sharper detail, precise edits, faster.

Lyria 3.5

Google DeepMind

Google's best-sounding music model. Full songs with structure, vocals, and lyrics.

MiniMax H3 Max Turbo

fal.ai

The fast tier of fal's post-trained H3. Twice the speed, nearly the same look.

Inworld TTS 2

Inworld

Ninety-five voices, 100+ languages. Speech that starts in under a second.