01
Choose the source
Begin with text, animate a frame, direct a transition, or use references when the selected model supports them.
Frontier video direction
Text or image.21 ways to move.
Start from a prompt, animate one image, direct first and last frames, or guide a shot with references. Change the model—not your production process.
Every model animates images Setup stays ready until you generate Results saved to your library
Video Lab / Model stage
21 onlineLoading Veo 3.1 sample
Official sample
Original model creatorGoogle
Original model creatorGoogleSourced model output / Lens 01
Veo 3.1
Director note
Describe the subject, action, camera move, lighting, and sound as separate beats in one concise shot direction.
Sample credited to official sample. Results vary with prompt, settings, and provider updates.
Duration4, 6, or 8 secondsResolution720p, 1080p, or 4KFrame16:9 or 9:16AudioOptional generated audio
From the community
Published Video Lab takes.
Open a real generation, then Generate this in Video Lab.
A better control surface
Model choice without workflow chaos.
Every frontier model has a different motion language. Video Lab keeps prompts, source images, frame roles, format, duration, resolution, audio, and cost in one consistent directing room.
02
Lock the setup
Set the model, duration, format, and controls before any generation is submitted.
03
Keep the take
Completed shots and their settings stay together in your history and Clip Studio Library.
The model slate
21 distinct creative lenses.
Official sample01GoogleVeo 3.1A premium text-and-image video model for cinematic shots, guided frames, references, optional audio, and up to 4K output.Text promptAnimate imageFirst → lastUp to 3 reference imagesSee model details
Official sample02GoogleVeo 3.1 FastA faster, lower-cost Veo 3.1 SKU for cinematic shots, guided frames, references, optional audio, and up to 4K.Text promptAnimate imageFirst → lastUp to 3 reference imagesSee model details
Official sample03
ByteDanceSeedance 2.5A next-generation text-and-image video model with native audio, 30-second shots, first/last frames, and deep reference control.Text promptAnimate imageFirst → lastUp to 30 reference imagesSee model details
Independent test04ByteDanceSeedance 2.0A flexible text-and-image model with first/last-frame direction, up to nine references, broad formats, and optional audio.Text promptAnimate imageFirst → lastUp to 9 reference imagesSee model details
Official sample05
KuaishouKling 3 ProA controllable text-and-image model for expressive movement, first/last frames, subject references, and optional audio.Text promptAnimate imageFirst → lastUp to 3 subject elementsSee model details
Official sample06KuaishouKling O3Kling’s newer O3 Standard generation for realistic motion, optional audio, intelligent multi-shot, first-to-last frames, and references.Text promptAnimate imageFirst → lastUp to 4 reference imagesSee model details
Official sample07
AlibabaHappy Horse 1.1A native-audio text-and-image model with source-image animation, multi-image references, and broad social formats.Text promptAnimate imageUp to 9 reference imagesSee model details
Independent test08GoogleGemini Omni FlashA fast multimodal 1.1 model for text, image animation, first/last frames, and references with native audio up to 4K.Text promptAnimate imageFirst → lastUp to 9 reference imagesSee model details
Official sample09AlibabaWan 3.0A longer 1080p one-pass text-and-image model with optional native audio, first/last frames, and up to ten references.Text promptAnimate imageFirst → lastUp to 10 reference imagesSee model details
Official sample10AlibabaWan 3.0 PrimeA faster Wan 3.0 SKU for 2–30 second 1080p shots with optional native audio, first/last frames, and up to ten references.Text promptAnimate imageFirst → lastUp to 10 reference imagesSee model details
Official sample11AlibabaWan 2.7A practical text-and-image generator with first/last frames, multi-image references, 1080p, and repeatable controls.Text promptAnimate imageFirst → lastUp to 9 reference imagesSee model details
Official sample12
MiniMaxMiniMax H3 Maxfal’s post-trained MiniMax H3 variant for native-audio clips, image animation, first/last frames, and references at 480p, 768p, or 1080p.Text promptAnimate imageFirst → lastUp to 9 reference imagesSee model details
Official sample13MiniMaxMiniMax H3A high-detail native-audio model with first/last frames, references, and 480p through 4K output.Text promptAnimate imageFirst → lastUp to 9 reference imagesSee model details
Independent test14
xAIGrok Imagine Video 1.5A flexible native-audio model for text, single-image animation, and multi-reference clips from 480p to 1080p.Text promptAnimate imageUp to 7 reference imagesSee model details
Official sample15
Black Forest LabsFLUX 3Black Forest Labs’ frontier video model for text, single-image motion, first-to-last frames, optional native audio, and clips up to 20 seconds.Text promptAnimate imageFirst → lastSee model details
Official sample16
PixVersePixVerse V6A broad text-and-image toolkit with image animation, first/last transitions, four resolutions, and style controls.Text promptAnimate imageFirst → lastSee model details
Official sample17
LightricksLTX 2.3 ProA production-oriented text-and-image model with first/last frames, 1080p to 4K output, and frame-rate controls.Text promptAnimate imageFirst → lastSee model details
Official sample18KuaishouKling O3 ProKling O3 Pro for realistic motion, optional audio, intelligent multi-shot, first-to-last frames, and references.Text promptAnimate imageFirst → lastUp to 4 reference imagesSee model details
Official sample19LightricksLTX 2.5 ProLightricks’ newer Pro SKU for 720p and 1080p clips with first/last frames, optional audio, and 24, 25, or 50 fps.Text promptAnimate imageFirst → lastSee model details
Official sample20
OpenAISora 2OpenAI Sora 2 for 720p clips with always-on native audio from text or a starting image.Text promptAnimate imageSee model details
Official sample21OpenAISora 2 ProOpenAI Sora 2 Pro for 720p or true 1080p clips with always-on native audio from text or a starting image.Text promptAnimate imageSee model detailsEditorial selection · Last reviewed 11 September 2026
Before the first take
Questions from the directing room.
What is Clip Studio Video Lab?
Video Lab is one workspace for generating videos with text prompts or source images across 21 curated frontier models. Choose a model, add a prompt or image, generate, and keep the completed clip in your Clip Studio Library.
Are these the 21 most popular AI video models?
No objective cross-provider popularity ranking exists. This is an editorial selection of 21 frontier video models chosen for useful differences in quality, speed, control, format support, audio, and resolution.
How is Video Lab priced?
Each plan includes a monthly generation limit. Model, duration, resolution, and other settings use that allowance. Upgrade anytime for a higher monthly limit.
Can I compare the same prompt across models?
Yes. Generate with one model at a time, then use Try another model to preserve your prompt, compatible settings, and available source images while you select a different model.
Which image-to-video workflows are supported?
Every Video Lab model can animate a starting image. Veo, Seedance, Kling, Wan, MiniMax, Gemini, PixVerse, LTX, and FLUX 3 also support first-to-last-frame direction, while Veo, Seedance, Happy Horse, Gemini, Wan, MiniMax H3, MiniMax H3 Max, Kling O3, Kling O3 Pro, and Grok offer multi-image reference workflows. Kling 3 Pro uses subject reference elements inside its image workflow. Sora 2 and Sora 2 Pro animate a starting image; they do not take first-to-last frames or references.
Bring the brief.Choose the motion.
Bring a prompt or source image, choose the right model, and generate.
Enter Video Lab