Alibaba launches Wan3.0 AI video model in public beta with 30-second clips and multimodal input support
AI video model releases
Alibaba launches Wan3.0 AI video model in public beta with 30-second clips and multimodal input support
Alibaba released a public beta of Wan3.0, an AI video generation model, accessible via Alibaba Cloud's Model Studio and Qwen Cloud platforms.
Wan3.0 generates clips up to 30 seconds long, double the 15-second limit of its predecessor Wan2.7-Video, and longer than the few-to-15-second outputs typical of mainstream AI video generators.
The model has an intelligent duration feature that recommends optimal clip length from a prompt and supports video extension to lengthen narrative timelines.
Wan3.0 accepts multimodal inputs including text, image, video, audio, web pages, PDFs, and PowerPoint files, and aims to reduce visual drifting by keeping faces, micro-expressions, multilingual voice output, and UI/motion graphics stable.
Alibaba positions Wan3.0 for filmmaking, short dramas, marketing and educational videos, and simulation training data for self-driving cars and robotics; the Wan series debuted in July 2023.
cho.sh