Alibaba's current video model — now on Cdance
Wan 3.0: Alibaba's Current AI Video Generator
Wan video generation starts at 3 credits per second; the total updates with model and resolution.
What's New in Wan 3.0
Wan 3.0 advances Alibaba's open video model line with sharper 1080p output, deeper bilingual comprehension, and motion quality that rivals closed-source flagships.
Crisp 1080p Generation
Wan 3.0 renders native 1080p video with clean textures, accurate colors, and fine detail that survives compression — quality that stands up on big screens, not just phone feeds.
Bilingual Prompt Mastery
Trained by Alibaba's Tongyi team, Wan understands Chinese and English prompts equally well — including culturally specific subjects, idioms, and aesthetics that Western models often miss.
Alibaba Model Lineup
As the flagship open-weights video model, Wan benefits from a massive community ecosystem of fine-tunes, LoRAs, and workflows — and Cdance brings the best of it to a one-click interface.
Text & Image to Video
Text & Image to Video is available through the supported controls shown in the generator.
Why Creators Choose Wan Video
Wan combines high-resolution output, bilingual prompt support, native audio, and model-specific generation controls.
Wan AI Video Technical Capabilities
Six core capabilities of Alibaba's Wan 3.0 video model.
Native 1080p Output
Generate full-HD video directly — no upscaling tricks. Wan 3.0 produces sharp, detailed 1080p frames suitable for professional editing and broadcast-adjacent work.
Chinese & English Prompts
Write prompts in either language — or both. Wan's bilingual training data delivers faithful interpretation of nuanced descriptions in Chinese and English alike.
Text-to-Video Generation
From single sentences to detailed shot descriptions, Wan 3.0 turns text into coherent, well-composed video with strong subject and action adherence.
Image-to-Video Animation
Animate product photos, illustrations, and portraits while preserving identity and style. Ideal for e-commerce, marketing assets, and creative projects.
Rich Motion & Style Range
Photorealism, anime, ink-wash painting, 3D animation and more — Wan's diverse training gives it one of the widest stylistic ranges of any video model.
Current API Availability
Wan's weights and research are public, driving rapid community improvement. New capabilities from the open ecosystem reach Cdance users faster than closed-model update cycles.
Wan AI Video Frequently Asked Questions
Common questions about Alibaba's Wan 3.0 and Wan 2.7 video models.
Generate with Wan 3.0 Today
Choose the model, configure supported settings, and review the exact total before generating.