NightCafe Logo
Create
Anonymous User

Wan 2.1

Video· ·14K+ creations·1 comment
A text-to-video & image-to-video model by Tongyi Lab of Alibaba.

Wan 2.1, developed by Alibaba’s Tongyi Lab, is an open-source text-to-video and image-to-video AI model launched in February 2025. Built on a 3D Causal Variational Autoencoder (Wan-VAE) and Diffusion Transformer (DiT) framework, it excels in creating realistic, high-definition videos with smooth motion, accurate physics, and bilingual (Chinese/English) text rendering.

On NightCafe, Wan 2.1 produces cinematic, photorealistic, or animated clips (up to 5 seconds) for marketing, storytelling, or art, supporting diverse styles and complex movements. Its lightweight T2V-1.3B variant runs on consumer GPUs, democratizing access for creators.

Topping the VBench leaderboard with an 84.7% score, Wan 2.1 offers NightCafe users a powerful, versatile tool for professional-grade video creation, enhanced by editing and audio generation capabilities.


Recent creations by the community

See what people have made with this model

0
2
Malfunction
2
3
Eerie Black Cat Sentinel in Gothic Cemetery
0
2
Reverent Warrior in Ancient Forest, Dramatic Lighting
0
3
Lion with Luminous Mane at Sunset: Digital Painting
0
2
Lion with Luminous Mane at Sunset, Hyperrealistic Style
4

Frequently asked questions

What is Wan 2.1?

Wan 2.1 is Alibaba’s open-source AI model, generating high-quality videos from text or images with realistic motion and text rendering.

What types of videos can Wan 2.1 create?

It produces photorealistic, cinematic, or animated clips (up to 6 seconds) with smooth transitions, ideal for marketing, art, or educational content.

How does Wan 2.1 handle text in videos?

It’s the first model to render both Chinese and English text accurately within videos, perfect for titles, captions, or branding.

Is Wan 2.1 accessible for beginners?

Yes, its intuitive interface on NightCafe allows beginners to create professional videos using simple, descriptive prompts.

What are Wan 2.1’s limitations?

Complex multi-object prompts may require refinement, and advanced features like Motion Brush are not yet supported.


Comments

Anonymous User
PRO

The best model I've used so far on NightCafe, produced an excellent animation of a martial arts battle and followed the prompt to a tee.

2 monthsReply

🤖 NightCafe

🔒 Secure Payments

PCI compliant payments powered by Paddle and PayPal and Shopify

PayPalShopifyShopify
MastercardVisaAmerican ExpressApple PayGoogle Pay

Copyright © NightCafe Studio Pty LtdCommunity Standards Acknowledgements Privacy Policy Terms of Service Refund Policy Sitemap