跳到正文
@alibaba_cloud· @alibaba_cloud · X·· 2026-09-01AI 评分24
AI 导读

阿里云在舞台上完整演示了从单条提示词到成片的视觉 AI 流水线:Qwen-Image 3.0 负责逐帧绘制,支持 12 种语言的文字、人物与版式;Wan 3.0 将画面生成为 30 秒带声音、不漂移的视频,为全新发布;WonderClip 用智能体跑通从剧本到终剪的全流程,达到工作室规模。活动于 9 月 4 日在曼谷举行。

正文

From a single prompt to a finished film. No studio. No crew. Just AI. We're walking through the full visual AI pipeline live on stage:

🎨 Qwen-Image 3.0: draws every frame right. Text, people, layouts in 12 languages.

🎥 Wan 3.0: animates those frames into 30-second, sound-on footage that never drifts. Brand new.

🎬 WonderClip: agents run the whole line, script to final cut, at studio scale. One pipeline. Concept to polished output.

Watch us build something real, right there on stage. 

📍 Sep 4, Bangkok — Register → https://t.co/NPB84yvra2

来源:@alibaba_cloud · x.com