WAN 3.0

WAN 3.0 is the latest generation of Alibaba's WAN video model. fal describes it as delivering enhanced motion smoothness, superior scene fidelity, and greater visual coherence. In Masonry it generates video from text or from a start image, with an optional end frame and generated audio.

Any whole-second duration from 2 to 30 seconds, with 5 seconds as the default. That makes it a fit for long single takes that would otherwise need several clips joined together.

WAN 3.0 outputs 480p, 720p, or 1080p, and defaults to 1080p. Use 480p or 720p to test an idea for less, then render the final version at 1080p.

16:9, 9:16, 1:1, 4:3, and 3:4. A prompt-only clip defaults to 16:9, and image-to-video follows the start image's framing unless you choose a ratio.

Yes. Image-to-video takes a start image as the first frame of the clip. Describe the motion and camera move rather than re-describing what the image shows.

Yes. You can add an optional end image that the clip finishes on. It requires a start image, so it is available on image-to-video only.

Yes. Audio is generated with the video and is on by default. Describe the sound you want in the prompt, or turn audio off for a silent clip.

Prompt expansion rewrites your prompt before generation and is on by default. fal notes that turning it off can save roughly 20 to 60 seconds but is likely to reduce quality, so leave it on unless you need the prompt followed literally or faster.

You can set a seed to make a generation repeatable, which helps when you want to change one detail at a time or rerun a test at a higher resolution.

Long, continuous takes that need smooth, coherent motion, such as product walkthroughs, slow reveals, and atmospheric brand films, especially at 1080p.

Masonry bills in credits by resolution and clip length, and the editor shows the exact cost before you generate.

fal marks both WAN 3.0 endpoints for commercial use. Review the current fal and Alibaba terms for your project, particularly for trademarks or a real person's likeness.

Open a Masonry canvas, pick WAN 3.0 from the model selector, and describe the video you need: a product shot, an ad creative, a social post. Masonry generates it, then you refine, edit, and combine WAN 3.0 with other models in one workspace.

New Masonry accounts get 50 free credits to try GPT Image 2.5 Flare. WAN 3.0 is included in every paid plan, and the editor shows the exact credit cost before you generate.

For a long take, describe how the shot develops over time, from the opening frame to the final one, including the camera path and any change in light or action. Name the sound you want next to the moment it belongs to. For image-to-video, describe the motion rather than the image, and when you add an end frame, describe what connects the two. Leave prompt expansion on for the best results, and turn it off only when you need the prompt followed more literally or faster.

WAN 3.0 is built by WAN Video. Inside Masonry it runs alongside 50+ image and video models, so your team can pick the right one for each brief without switching tools.