Flow - 4x faster than typing
Mini toolsWan Dancer 14B
By Alibaba
Model family: Wan
Wan Dancer 14B is a 14B parameter video model for minute-scale, coherent music-to-dance video synthesis. Given a reference image, a music track, and a text prompt describing dance style, it generates a stable, temporally coherent dance video at 720p and 30fps exceeding one minute, surpassing the roughly 20-second limit of most diffusion video models. It uses a hierarchical two-stage pipeline: global keyframe planning for long-range coherence followed by local temporal refinement for motion detail. Key techniques include dynamic frame-rate adaptation via time-mapped RoPE embeddings, an optical-flow-based loss for motion continuity, and motion-speed control to preserve fidelity during rapid movement. It supports five dance genres: Chinese classical dance, K-pop, street dance, Latin dance, and tap dance. Released under the Apache 2.0 license with open weights and code.
New
Video
Overview
Wan Dancer 14B is an image-to-video and audio-to-video model that generates minute-scale, coherent dance videos from a reference image, a music track, and a dance style prompt. It produces stable 720p, 30fps dance videos exceeding one minute, overcoming the roughly 20-second limit of typical diffusion video models via hierarchical global keyframe planning and local temporal refinement. It supports five dance genres: Chinese classical, K-pop, street, Latin, and tap dance.
About Alibaba
Chinese e-commerce and cloud leader behind Taobao, Tmall, and Alipay.
View Company Profile
Last updated: July 14, 2026
