Start-frame preservation
“Slow push toward the subject as wind moves the coat and background foliage; keep the original face and composition.”
Check identity, framing, texture stability, and whether motion changes unrequested details.
Free Wan 2.2 AI Video Generator
Create cinematic AI videos with Wan 2.2 in Magic Hour. Use it for flexible text-to-video, image-to-video, controllable aesthetics, and open workflow experimentation with strong visual quality.
Rustic kitchen
View prompt
A grandmother kneads bread dough on a flour-covered wooden table in a rustic kitchen, sunlight through a small window, slow push-in
Fantasy jungle
View prompt
A glowing blue river winds through a giant fantasy jungle, bioluminescent plants pulse with light as a small boat drifts through, the camera flies forward
City street
View prompt
A busy city street at golden hour, people crossing as a cyclist rides past the camera, warm light reflecting off shop windows
Cyberpunk VR city
View prompt
A person wearing a VR headset stands still as a neon cyberpunk city builds itself around them block by block, the camera slowly circles
Hallway portrait
View prompt
A single man in a dark suit stands alone in a dim hallway, slowly turns his head to face the camera and smiles slightly, static camera, warm practical light, shallow depth of field
Wildflowers
View prompt
Wildflowers sway in a gentle breeze on a hillside and a butterfly lands on a daisy, soft sunlight, camera low among the flowers
Speaker
View prompt
A woman speaks confidently into a microphone on a small stage, gesturing to the audience, spotlight, slow push-in
Skate park
View prompt
A skateboarder drops into a halfpipe and launches into an aerial spin at sunset, slow motion at the peak of the jump
Elegant portrait
View prompt
Elegant portrait of a woman in a black evening gown slowly turning toward the camera in a candlelit ballroom, soft focus background
How to evaluate Wan 2.2
A practical decision guide based on Magic Hour's current product support and the questions creators ask when comparing models.
Best fit on Magic Hour
Visual-first text-to-video and image-to-video work where cinematic composition, lighting, and open-model flexibility are the priority.
Consider another model when
Choose a model with native audio when dialogue or synchronized sound must arrive in the same generation.
Three useful first-run tests
Use the prompts below to test the failure modes that matter before committing to a larger batch.
Fast physical motion
“A dancer spins, drops to one knee, rises, and crosses frame while the camera tracks sideways.”
Watch limbs, body proportions, floor contact, motion direction, and camera smoothness.
Cinematic lighting change
“A storm passes over a coastal road as warm sunset light breaks through and reflects across the wet pavement.”
Check lighting continuity, reflections, weather motion, and preservation of the scene layout.
These prompts are evaluation recipes, not controlled benchmark results. Run the same prompt and settings across models before drawing a conclusion.
Trusted by teams at
How To Generate Videos with Wan 2.2
Three simple steps to create AI videos with Wan 2.2 in Magic Hour
- 1
Type a Prompt or Upload an Image
Start with a text prompt for text-to-video, or upload a still image if you want image-to-video control. - 2
Customize Video
Adjust aspect ratio, duration, and settings based on whether you want a cleaner cinematic shot, a reference-led animation, or a more exploratory generation. - 3
Generate Your Video
Click "Generate Video" to start. Once it's ready, download the clip and compare it against other leading models if you want the best trade-off between control, speed, and polish.
Why Creators Choose Wan 2.2
Where Wan 2.2 really shines: cinematic visual control, flexible open workflows, strong image-to-video performance, and the ability to get high-quality results without being locked into a purely closed model stack.
Cinematic control
Wan 2.2 is especially useful when you want stronger control over lighting, composition, color tone, and general film-like aesthetics.
Flexible image-to-video workflows
It is a strong fit when you want to start from a still image and push it into a controlled, high-quality motion result.
Open workflow flexibility
A major reason creators use Wan 2.2 is that it works well in open-source and customizable production setups rather than only inside a closed model surface.
No Editing Skills Required
Turn your ideas into polished outputs quickly — no complex editing workflow needed. Just use a clear prompt, a reference if helpful, and iterate.
Thousands of Templates
Make your photos even better and faster with our high-quality templates.
Browse Wan 2.2 TemplatesHow Wan 2.2 Compares to Other Video Models
Wan 2.2 THIS MODEL | Veo 3.1 | Kling 3.0 | MiniMax H3 | LTX-2.5 | Seedance 2.5 | |
|---|---|---|---|---|---|---|
Best for | Open flexible video generation, cinematic control, strong text-to-video and image-to-video workflows | Premium realism, polish, dialogue, and strong prompt adherence | Multi-shot storytelling, character references, structured cinematic control | Reference-driven clips with native audio, character consistency, longer short-form | Fast iteration, synced audio, uploaded audio, clips up to 60s | Cinematic continuity, structured references, narrative short-form |
Max output | 1080p | 1080p | 4K | 1080p | 1080p | 1080p |
Real-human image to video | Good | Better suited | Better suited | Better suited | Good with the right source image | Good |
Viral clips | Good | Good | Strong | Strong | Useful for rapid drafts | Strong |
Multi-scene storytelling | Medium to strong | Strong | Excellent | Medium to strong | Long clips; test continuity | Strong |
Surreal / imaginative concepts | Good | Strong | Good | Good | Prompt-led generation | Strong |
Native audio | No in current Magic Hour workflow | Yes | Yes | Yes | Yes | Yes |
Template fit | Excellent | Good | Good | Good | Good | Good |
What Wan 2.2 is best at
Where Wan 2.2 really shines: cinematic visual control, flexible open workflows, strong image-to-video performance, and the ability to get high-quality results without being locked into a purely closed model stack.
Cinematic aesthetics
Especially strong when lighting, composition, and overall mood matter.
Image-to-video quality
A very solid choice for turning a still frame into a controlled moving shot.
Open-source production use
Useful for teams that want inspectability, customization, and integration flexibility.
Strong all-around visual generation
Good at balancing aesthetics, motion, and prompt readability.
Template-friendly workflows
Works well when paired with repeatable presets and reference-led setups.
A strong default open model
A practical option when you want something more controllable than a black-box premium model.
Wan 2.2 — Model Card
Key specs, capabilities, and limitations.
Cost
24–72 credits/sec depending on resolution
Resolution
480p, 720p, 1080p
Aspect ratios
Common portrait and landscape ratios; some workflows follow source image aspect ratio
Max duration
Up to 15s
Audio
No native audio in current Magic Hour workflow
Plans
Free / Creator / Pro / Business
Limitation
More workflow-sensitive than closed premium models; best results often come from stronger prompting or references
Frequently asked questions
Wan 2.2 is Alibaba's open video generation family built for cinematic text-to-video and image-to-video workflows. It adds MoE-based architecture, stronger aesthetic control, and multiple task variants inside the same broader model family.
Wan 2.2 is best for flexible video generation, cinematic-looking shots, strong image-to-video workflows, and open-source production setups where control matters.
Inside Magic Hour, Wan 2.2 supports 480p, 720p, and 1080p output at 16fps, 9:16, 16:9, and 1:1 aspect ratios, and durations up to 15 seconds. Free users get 480p; 720p and 1080p need a paid plan.
Not in the standard Wan 2.2 text-to-video and image-to-video workflows. Treat it as a visual-first model unless you are specifically using another audio-driven Wan workflow.
Yes. Wan 2.2 has a dedicated image-to-video model and is one of the stronger open options for turning a still image into a controlled cinematic clip.
Wan 2.2 costs 24 credits per second at 480p, 48 at 720p, and 72 at 1080p. 480p is available on all plans; higher resolutions require a paid plan.
Yes. Wan 2.2 was released publicly with weights and inference code, which is a major reason it matters for developers and production teams.
Wan 2.2 accepts real-human source images. Test identity and motion on the actual image; compare a premium model when output polish matters most.
Wan 2.2 supports 9:16, 16:9, and 1:1 inside Magic Hour.
Magic Hour lets you use Wan 2.2 in a simpler product workflow, compare it against Veo, Kling, MiniMax, LTX, and Seedance, and pair it with templates instead of managing a raw open-source stack yourself.
Choose Wan 2.2 for visual-first motion and action, including 480p on the free plan. Wan 3.0 adds native audio, end frames, and clips up to 30 seconds. Seedance 2.5 is the premium realism recommendation; Kling 3.0 adds multi-shot control; MiniMax H3 offers references and audio; LTX-2.5 supports fast generation up to 60 seconds.
Paid subscriptions, credit-pack purchases, and separately agreed paid usage-based billing arrangements include commercial-use rights. Users without one of these paid arrangements are limited to personal, non-commercial use. Outputs made with Wan 2.2 on Magic Hour can be used commercially, subject to our Terms.
Related Models
Compare models and create from text, images, or clips with Magic Hour's AI video generator.