Home/Explore/Wan 2.2 Video Models/wavespeed-ai/wan-2.2/i2v-480p-ultra-fast

image-to-video

wavespeed-ai/wan-2.2/i2v-480p-ultra-fast

Generate unlimited AI videos with Wan 2.2 A14B image to video ultra fast model.

Hint: You can drag and drop a file or click to upload

preview

Hint: You can drag and drop a file or click to upload

Idle

Your request will cost $0.05 per run.

For $1 you can run this model approximately 20 times.

One more thing:

ExamplesView all

README

Wan-2.2-i2v-480p-ultra-fast

wan-2.2/i2v-480p-ultra-fast is the ultra-fast, optimized version of Wan 2.2, designed for image-to-video (I2V) generation at 480p resolution. It leverages an innovative Mixture of Experts (MoE) architecture, combining high-noise and low-noise experts across denoising timesteps, enabling faster rendering while maintaining cinematic quality.

Why it looks great

  • Cinematic Aesthetics – integrates professional film standards, controlling lighting, color, and composition.
  • Natural Motion – handles large-scale, complex movements with smoothness and controllability.
  • Semantic Precision – understands complex scenes and multi-object interactions, staying true to creative intent.
  • Ultra-fast Performance – optimized pipeline ensures faster turnaround without sacrificing visual fidelity.

Limits and Performance

  • Resolution: 480p
  • Duration: 5s or 8s

Pricing

Video LengthCost (USD)
5 seconds$0.05
8 seconds$0.08

Billing Rules:

  • Minimum charge: 5 seconds
  • Per-second rate: $0.01
  • Total cost = video length × $0.01

How to Use

  1. Upload image – the keyframe (first frame) or reference image for video generation.
  2. (Optional) Upload last_image – to guide video generation.
  3. Enter prompt – describe scene, camera motion, style, or action.
  4. (Optional) Set negative_prompt – exclude unwanted elements.
  5. Choose duration – default 5s, adjustable for longer clips.
  6. Set seed – use -1 for randomness, or fix a number for reproducibility.
  7. Click Run – rendering begins, optimized for ultra-fast output.

Pro Tips

  • Combine image + prompt for stronger consistency.
  • Keep prompts detailed (motion, mood, lighting) for cinematic effect.
  • Fix seed for reproducibility, change seed for variety.
  • Use short clips for best speed-to-quality balance.

Note

When uploading an image, please check if the URL is accessible.

  • If the URL is invalid or blocked, the model will not be able to load the image.
  • A valid image URL will display a preview in the interface.