Qwen /

Wan3.0-Video

70 DZD/ second

Wan3.0-Video is an all-in-one multimodal AI model launched by Alibaba's Tongyi Lab, integrating video generation and precise editing into a single unified architecture. Replaced the previous task-specific versions, it supports native 30-second clips at up to 1080p with simultaneous native audio-visual generation in one single pass. Beyond standard text and images, its signature Omni-Reference capability allows it to parse up to 20 assets—including documents (PDFs, PPTs) and webpages—to directly drive character consistency, movement replication, and precise pixel-level interactive editing.