Latest Model

Wan 3.0 AI Video Generator

30-Second Cinematic Video with Native Audio

Wan 3.0 is Alibaba's newest video model: a single pass produces up to 30 seconds of 1080p footage with synchronized sound. Guide it with text, a start and end frame, or up to 10 reference images.

30 Seconds
1080p Ultra HD
Native Audio
Reference to Video
Start Creating
30sMax Duration
1080pUltra HD
AudioNative Sound

Wan 3.0 Video Generator

Wan 3.0 typically takes 3-5 minutes; 30-second clips at 1080p can take longer.

Wan Video Generator

0/2000
0/1000

Recent Generations

Example Videos

See what you can create - Sign in to start generating your own videos

16:9
Example #1

Cinematic single take: a neon-lit Tokyo side street at night in heavy rain, a woman in a yellow raincoat walks toward camera holding a transparent umbrella, reflections shimmer on wet asphalt, slow dolly-in, shallow depth of field, ambient rain and distant traffic sounds

16:9
Example #2

Aerial drone shot rising over turquoise ocean waves breaking on a white sand beach at golden hour, palm shadows stretch across the sand, camera tilts up to reveal the sun on the horizon, gentle wave sounds and wind

16:9
Example #3

Close-up handheld shot in a busy restaurant kitchen: a chef flips vegetables in a flaming wok, sparks and steam rise, warm tungsten light, quick rack focus to the plated dish, sizzling audio and kitchen chatter

Technical Specifications

What Wan 3.0 delivers on Fast Wan

480p / 720p / 1080p

Resolution

Three quality tiers, up to Full HD

5-30 seconds

Duration

Single-pass generation, no stitching

Native

Audio

Sound generated in the same pass

First & last frame

Frame Control

Pin where the clip starts and ends

16:9, 9:16, 1:1, 4:3, 3:4

Aspect Ratios

All major platform formats

Text, Image, Reference

Generation Modes

Up to 10 reference images

Wan 3.0 Key Features

What changed from Wan 2.7 and why it matters for your videos

30-Second Single Pass

Wan 3.0 doubles the length ceiling of Wan 2.7. A full 30-second clip is generated in one pass, so motion, lighting and characters stay consistent without stitching shorter segments together.

Native Audio Generation

Ambient sound, effects and speech are produced alongside the picture. Describe the soundscape in your prompt and the audio track lands in sync with the visuals.

First & Last Frame Control

In image-to-video mode you can supply both a start frame and an end frame. Wan 3.0 animates a smooth transition between them - ideal for product reveals and seamless loops.

Reference-to-Video

Upload up to 10 reference images and cite them in the prompt ("the character in Image 1"). Wan 3.0 keeps the same subject, outfit and style across the whole clip.

1080p Ultra HD Output

Choose 480p for quick drafts, 720p Pro for everyday content, or 1080p Ultra for final delivery. Credits scale with resolution, so you only pay for the quality you need.

Commercial-Ready

Use Wan 3.0 output in ads, product videos, social campaigns and client work. Generations are private to your account and downloadable as MP4.

Perfect For

Where Wan 3.0's longer, audio-ready clips make the biggest difference

Short-Form Social Video

Produce complete 15-30 second TikTok, Reels and Shorts clips with sound in one generation - no separate audio pass needed.

01

Product & Brand Ads

Lock the opening and closing frame around your product shot, then let Wan 3.0 fill the motion in between for polished ad spots.

02

Character-Consistent Stories

Use reference images to keep the same character across scenes for storytelling, explainer series and episodic content.

03

Concept & Previsualization

Draft 480p versions of a scene in minutes to test camera moves and pacing before committing to a 1080p render.

04

Wan 3.0 vs Wan 2.7 vs Wan 2.6

How the newest Wan model compares to the previous two generations on Fast Wan

Feature
LatestWan 3.0
Wan 2.7Wan 2.6
Max Duration30 seconds15 seconds15 seconds
Max Resolution1080p1080p1080p
Reference ImagesUp to 10Up to 5Not supported
First & Last FrameStart frame only
720p Price40 credits/s40 credits/s40 credits/s

How to Use Wan 3.0

01

Choose Your Mode

Pick Text-to-Video for a prompt-only clip, Image-to-Video to animate a photo (optionally with an end frame), or Reference-to-Video to keep a specific subject.

02

Write Your Prompt

Describe the shot as one continuous take: subject, action, camera movement, lighting. Add the sounds you expect so the audio track matches.

03

Set Duration & Quality

Choose 5-30 seconds, a quality tier (480p / 720p / 1080p) and the aspect ratio for your platform. Shorter drafts are the cheapest way to iterate.

04

Generate & Download

Click generate and wait 3-5 minutes. Preview the result with audio, then download the MP4 or send it to the upscaler.

Wan 3.0 Frequently Asked Questions

Common questions about generating videos with Wan 3.0

Wan 3.0 is the third-generation video model from Alibaba's Wan family, released in August 2026. It generates up to 30 seconds of video at up to 1080p in a single pass, produces audio alongside the picture, and accepts text, start/end frames and reference images as input.
The headline change is length: 30 seconds versus 15. Wan 3.0 also raises the reference-image limit from 5 to 10 and improves consistency of characters and scene layout across the longer clip. Pricing per second at 720p is unchanged.
Yes. Audio is generated in the same pass as the video. Describe the ambience, effects or speech you want in the prompt and it is rendered in sync with the visuals. You can still add or replace audio later with the Add Audio tool.
Credits are billed per second of output: 20 credits/s at 480p, 40 credits/s at 720p and 80 credits/s at 1080p. A 10-second 720p clip costs 400 credits; a 30-second 1080p clip costs 2,400 credits. Wan 3.0 Prime, the higher-fidelity tier, is 25/50/100 credits per second.
Prime is a premium tier of Wan 3.0 tuned for higher visual fidelity. It supports the same modes, durations and aspect ratios and costs about 25% more per second. Select it from the model list in the generator.
Upload 1 to 10 reference images and refer to them by position in your prompt, for example "the woman in Image 1 walks through the market in Image 2". Wan 3.0 preserves the subject's appearance and the scene's style throughout the clip.
Most 5-15 second clips finish in 3-5 minutes. 30-second clips and 1080p renders take longer. Your video keeps processing in the background, and you can find it under My Creations when it is ready.
Yes. Videos you generate on Fast Wan are yours to use in advertising, products and client work. Generations are private to your account by default.