Short cinematic concepts
Turn a detailed shot description into a polished visual-and-audio clip for previsualization, mood exploration, and creative pitching.
Independent tool overview
Veo 3.1 at a glance
Veo 3.1 is Google's current Veo video-generation family for short cinematic clips with native audio, portrait or landscape output, image guidance, frame control, and video extension.
Visit the official Veo 3.1 site ↗
- Current model
- Veo 3.1
- Original Veo 3 status
- Gemini API models retired June 30, 2026
- Clip length
- 4, 6, or 8 seconds per generation
- Resolution
- 720p, 1080p, and supported 4K output
- Aspect ratios
- 16:9 landscape and 9:16 portrait
- Audio
- Native sound, ambience, effects, and dialogue cues
- Access
- Google Flow, Gemini API, and Vertex AI
- Reviewed
- August 30, 2026
Overview
What Veo 3.1 is
Veo 3.1 is the current successor to Veo 3, Google's first broadly available video model with native synchronized audio. It creates 4-, 6-, or 8-second clips from text or images in landscape or portrait format, with 720p, 1080p, and—on supported variants—4K output. It can use first and last frames, reference images for a person, character, product, or style, and extensions of prior Veo clips, making it useful for concept shots, ads, social creative, storyboards, product visuals, and cinematic experiments.
The original Gemini API models `veo-3.0-generate-001` and `veo-3.0-fast-generate-001` were retired on June 30, 2026, so new integrations should use Veo 3.1 or Google's newer default video model where appropriate. Google now recommends Gemini Omni Flash for general conversational video generation and editing, while Veo 3.1 remains relevant for scene extension, explicit last-frame control, reference-image direction, and existing Veo workflows. Access and capabilities differ across Google Flow, the Gemini API, and Vertex AI, so users should confirm the active model, resolution, credit cost, and launch stage before each production workflow.
Use cases
Who Veo 3.1 is best for
The strongest fit depends on the job you need the product to complete, not the size of its feature list.
Product and character consistency
Use reference images to guide the appearance of a product, character, scene, or style across a generated shot.
Social and advertising variations
Generate portrait or landscape creative, test several prompts and variants, and use Fast or Lite models when iteration volume matters more than maximum quality.
Controlled transitions
Define both a starting and ending frame when a shot must move between two planned visual states.
Extending Veo scenes
Continue compatible Veo-generated clips in seven-second increments for longer sequences while preserving the preceding action and visual context.
Capabilities
Core Veo 3.1 features
Native video and audio generation
Creates visuals together with ambient sound, effects, music cues, and spoken dialogue described in the prompt.
Text-to-video
Generates a shot from a detailed scene description including subject, action, setting, camera, lighting, style, and audio direction.
Image-to-video
Animates a supplied starting image while following prompt guidance for motion, camera behavior, atmosphere, and sound.
First-and-last-frame control
Interpolates a shot between defined opening and closing images for more deliberate composition and transitions.
Reference images
Uses up to three images to guide the identity or appearance of a person, character, product, or visual ingredient in supported Veo 3.1 variants.
Video extension
Extends eligible Veo-generated videos by seven seconds at a time, up to 20 extensions and a combined maximum described as 148 seconds.
Portrait and landscape output
Supports 16:9 and 9:16 aspect ratios for widescreen, mobile, and social-video use cases.
Multiple quality and speed variants
Offers Lite, Fast, and Standard or Quality choices across access surfaces, trading cost and generation speed against resolution and output quality.
720p, 1080p, and 4K
Supports higher-resolution output on applicable models; 1080p and 4K generations require eight-second duration, while Lite tops out at 1080p.
Google Flow workflow
Provides a creative interface for prompting, ingredients, frames, extensions, scene building, asset management, and direct YouTube handoff.
Programmatic API access
Runs asynchronously through the Gemini API and is also available in Vertex AI with access, quota, model IDs, and launch stages that differ by platform.
SynthID watermarking
Embeds an invisible watermark in Veo-generated video so Google tools can help identify AI-generated content.
Safety and memorization checks
Screens prompts and outputs and applies checks intended to reduce harmful content, privacy, copyright, and memorization risks.
Process
How the Veo 3.1 workflow works
Step 1
Choose the access surface
Use Flow for hands-on creative work, the Gemini API for application workflows, or Vertex AI when Google Cloud governance, quota, and deployment controls are required.
Step 2
Pick the current model
Select Veo 3.1 Lite, Fast, or Quality based on feature compatibility, resolution, iteration speed, and cost; do not start new work on retired Veo 3.0 API IDs.
Step 3
Design one shot
Write a prompt with subject, action, environment, camera framing and movement, lighting, style, timing, dialogue, sound effects, and ambience rather than trying to direct an entire film at once.
Step 4
Add visual constraints
Supply a starting image, first and last frames, or up to three reference images when composition or subject identity must be more controlled.
Step 5
Generate and select
Create several candidates, inspect frame consistency, physics, text, faces, dialogue, lip sync, audio, and brand accuracy, then keep only outputs that pass review.
Step 6
Extend or edit the sequence
Use compatible extensions or Flow scene tools to build continuity, then finish timing, transitions, color, audio, captions, disclosure, and rights review in an editor.
Cost
Veo 3.1 pricing and free plan
Veo pricing depends on access method. Google Flow uses per-generation credits, while the Gemini API bills by successful output second with separate Lite, Fast, and Standard rates. Flow gives non-subscribers 50 daily credits; paid Google AI plans include monthly credits, and extra credits can be purchased in supported regions. Rates and credit costs can change, so verify the selected model in Flow or the official API pricing page before a large run.
Google Flow free credits
50 credits/day
Daily credits for non-subscribers to try Veo 3.1 Lite, Fast, and Quality generations in Flow.
- Unused free credits do not roll over
- Lite costs 10 credits per generation for non-Ultra users
- Fast costs 20 credits per generation for non-Ultra users
- Quality costs 100 credits per generation
- One request may create multiple separately charged generations
Veo 3.1 Lite API
$0.05/sec 720p · $0.08/sec 1080p
Lowest-cost Gemini API Veo variant for short video with audio.
- Approximately $0.40 for an 8-second 720p clip
- Approximately $0.64 for an 8-second 1080p clip
- 4K is not supported
- No Gemini API free tier
Veo 3.1 Fast API
$0.10–$0.30/sec
Faster paid API generation for higher-volume creative and application workflows.
- $0.10/sec at 720p
- $0.12/sec at 1080p
- $0.30/sec at 4K
- An 8-second clip costs about $0.80, $0.96, or $2.40 respectively
Veo 3.1 Standard API
$0.40/sec · $0.60/sec 4K
Highest-priced Gemini API Veo quality tier.
- $0.40/sec at 720p or 1080p
- $0.60/sec at 4K
- An 8-second clip costs about $3.20 or $4.80 at 4K
- Google charges only for successfully generated output when audio processing or safety blocks prevent completion
Pricing checked . Check current pricing at the source ↗
Assessment
Veo 3.1 strengths and limitations
Where it stands out
- Generates visuals and audio together instead of requiring a separate sound-design pass for every concept.
- First-and-last-frame control and reference images provide more direct composition guidance than text prompting alone.
- Portrait and landscape support makes the same model family useful for social, advertising, and cinematic formats.
- Lite, Fast, and Quality choices let teams separate cheap iteration from higher-cost final candidates.
- Veo extension can turn a successful short shot into a longer compatible sequence.
- Flow provides a creative interface while Gemini API and Vertex AI support programmatic and governed deployments.
- Published per-second API pricing makes cost modeling more transparent than credit-only systems.
- SynthID and safety checks provide a built-in provenance and risk-reduction layer.
- Google's broader image, video, Gemini, and Flow ecosystem supports an end-to-end concept workflow.
What to consider
- The original Veo 3.0 Gemini API models are retired; integrations that did not migrate by June 30, 2026 will fail and need a current model ID.
- Google now recommends Gemini Omni Flash as the default for general video generation, leaving Veo 3.1 positioned around specific controls and legacy workflows.
- Gemini API Veo 3.1 model IDs are documented as preview, so availability, behavior, quotas, and compatibility can change before a stable release.
- Each base generation is only four, six, or eight seconds, and higher resolutions or reference-image workflows generally require eight-second output.
- Video extension is limited to compatible Veo-generated clips, works at 720p, and depends on source assets retained or referenced within the platform's time window.
- Generated API videos are stored for two days, so applications must download and manage files promptly.
- Native dialogue and short spoken audio can still be incoherent or poorly synchronized; Google describes this as an active area of development.
- Safety filters and audio processing can block a generation even when a prompt appears acceptable, which affects predictable batch completion.
- Generation latency can range from roughly 11 seconds to six minutes during peak periods according to the API documentation.
- English is the fully supported prompt language; results from other languages may be inconsistent or unevaluated.
- Person-generation controls and allowed inputs vary by region, particularly in the EU, UK, Switzerland, and MENA.
- Visual continuity, object permanence, physics, hands, typography, logos, dialogue, and character identity still require inspection across every frame.
- Repeated generation can become expensive because many candidates may be needed for one usable shot, and Flow requests can produce multiple charged generations.
- Reference images, uploads, music, voices, likenesses, brands, and commercial use still require appropriate rights and consent; safety filters do not provide legal clearance.
- SynthID helps identify Google-generated media but is not a complete substitute for visible disclosure, provenance records, and publisher policies.
Compare
Veo 3.1 alternatives
The right alternative depends on the specific output, workflow, controls and budget your project requires.
Content Creator
Runway Gen-4.5
Runway Gen-4.5 is a strong creative-production alternative with its own motion, consistency, editing, and filmmaking workflow; compare controllability, audio, clip economics, and editing needs.
Explore Runway Gen-4.5 →Content Creator
Hailuo AI
Hailuo AI is worth comparing for fast text- and image-to-video iteration, especially when broad consumer access and lower-cost experimentation matter.
Explore Hailuo AI →Content Creator
LTX Studio
LTX Studio is better suited when the priority is planning and assembling a larger story with scripts, shots, characters, and timelines rather than calling one video model directly.
Explore LTX Studio →Questions
Veo 3.1 FAQs
Is Veo 3 still available?
The original `veo-3.0-generate-001` and Fast Gemini API models were retired on June 30, 2026. The Veo product continues as Veo 3.1 through Flow, Gemini API preview models, and supported Vertex AI model versions.
How much does Veo 3.1 cost?
Gemini API prices range from $0.05 per second for 720p Lite to $0.60 per second for 4K Standard. Google Flow uses credits, including 50 free daily credits for non-subscribers and monthly allocations with Google AI plans.
Does Veo 3.1 generate audio?
Yes. Audio is always on for the documented Veo 3.1 API variants and can include dialogue, ambience, sound effects, and other cues, though speech quality and synchronization can still fail.
How long are Veo 3.1 videos?
A base generation can be four, six, or eight seconds. Certain features and higher resolutions require eight seconds. Compatible Veo clips can be extended in seven-second increments under documented limits.
Can Veo 3.1 make vertical video?
Yes. It supports both 16:9 landscape and 9:16 portrait output, although feature compatibility can vary by model and access surface.
Can I keep a character or product consistent?
Veo 3.1 can use up to three reference images to guide a person, character, product, or style. This improves direction but does not guarantee perfect identity, logo, or detail consistency.
What is the difference between Veo 3.1 Lite, Fast, and Quality?
Lite is the cheapest, Fast prioritizes speed and volume, and Standard or Quality targets stronger output at a much higher cost. Resolution, extension, reference, and Flow feature support differ, so check the active model before generating.
Should developers use Veo 3.1 or Gemini Omni Flash?
Google recommends Gemini Omni Flash as the default for general video generation and conversational editing. Veo 3.1 remains useful when you need Veo scene extension, explicit first/last-frame control, reference-image direction, or compatibility with an existing Veo pipeline.
Are Veo videos watermarked?
Yes. Google says Veo-generated videos contain an invisible SynthID watermark. Flow may also apply or offer visible watermarking depending on plan and region.
Bottom line
Our Veo 3.1 verdict
Veo 3.1 remains one of Google's most capable shot-generation options when native audio and explicit creative controls matter. Reference images, first-and-last-frame interpolation, portrait output, extensions, and clear API rates make it practical for disciplined production experiments. The key is to treat the model as a shot generator, not a finished-film button: plan for multiple attempts, manual editing, rights review, audio cleanup, provenance, and model migration. New API work should use current Veo 3.1 IDs—or evaluate Gemini Omni Flash first—rather than the retired Veo 3.0 models this page originally covered.
Visit Veo 3.1 website ↗Similar Content Creator Tools
View all Content Creator AI tools →
Co-STORM
Co-STORM - Write Wikipedia-like articles from scratch based on AI search
Genspark AI Docs
Genspark AI Docs - an agentic creator allowing users to generate and edit a variety of document times via natural language prompts.
HeyGen Video Agent
HeyGen Video Agent - Create video content with scripts, actors, edits, and more
Marey by Moonvalley
Marey - A filmmaker-focused AI video model trained exclusively on licensed content that gives directors granular control over scenes
AI training for the future of work.
Get access to all our AI courses, hundreds of real-world AI use cases, live expert-led workshops, an exclusive network of AI early adopters, and more.
AI Courses
Get unlimited access to all of our current & upcoming industry-specific AI courses for the duration of your subscription.
Daily Guides
To keep up with the rapid pace of AI, our team publishes AI implementation guides daily. Our library contains 300+ practical use cases to automate real-world work.
Workshops
Join weekly, live, interactive sessions with industry leaders who are at the forefront of AI for hands-on implementation guidance and exclusive insights.
Community
Network with an exclusive community of AI-first professionals who are working smarter with AI. Learn how early adopters are using AI in their work and businesses.