Le modèle Wan 3.0 à grande vitesse : créez des vidéos de 30 secondes avec du son natif à partir de textes, d’images, de vidéos, de références audio, de documents et de pages web – génération nettement plus rapide.
Generate a video purely from a text prompt (up to 30 seconds).
Describe the scene, motion, and style you want in the video
2-30 seconds
480P : 20 gemmes/s, 720P : 40 gemmes/s, 1080P : 80 gemmes/s
Adaptive automatically matches the ratio of your input media and intent.
Ready to Create Your Wan 3.0 Prime Video
Wan 3.0 est le dernier modèle de génération vidéo AI tout-en-un d'Alibaba. Il transforme les invites de texte, les images, les vidéos, l'audio, les cadres clés, les documents et les pages Web en vidéos d'une durée pouvant aller jusqu'à 30 secondes avec un son synchronisé natif, à une résolution 480P, 720P ou 1080P.
Un modèle couvre le texte-vidéo, l'image-vidéo, les transitions de premier/dernier cadre, la génération de référence multimodale, l'édition vidéo et l'extension vidéo - avec une cohérence de référence parfaite en pixels qui garde les personnages, les vêtements, les accessoires et les environnements alignés avec votre matériel source tout au long du clip.
Start from text, images, videos, audio, keyframes, documents or even a webpage link. Wan 3.0 Prime reads all your source material and shapes scene, motion, sound and narrative around it.
Generate single clips up to 30 seconds long — enough room for context, action, transitions and a satisfying conclusion, without stitching fragments together.
Characters, clothing, props, environments and audio cues stay closely aligned with your references across changing actions and scenes — dependable for fashion films, character showcases and branded content.
Native audio generation pairs dialogue, music and environmental sound with the visuals, with natural timbre, regional speech expression and tight audiovisual synchronization.
Turn briefs, presentations, spreadsheets, PDFs and public webpages directly into dynamic video — perfect for lessons, demos, summaries and business content.
Complex action choreography, realistic physical interactions, detailed materials and nuanced performances — with first/last frame control for storyboard-precise transitions.
Pick Text-to-Video, Image-to-Video, First & Last Frame, or Reference Assets mode (images, videos, audios, document or webpage link).
Write a prompt (use @Image1/@Video1/@Audio1 tokens to point at reference assets), upload frames, or attach reference material.
Choose 2-30s duration, 480P/720P/1080P resolution, an aspect ratio, and whether to generate a synchronized audio track.
Click Generate and wait for processing. Preview the result and download your Wan 3.0 Prime video when ready.
Wan 3.0 Prime generates videos up to 30 seconds long with native synchronized audio from text prompts, first/last frame images, and multimodal references including up to 10 images, 5 videos, 5 audios, a document or a webpage link.
Wan 3.0 Prime outputs 480P, 720P or 1080P video with any whole-second duration from 2 to 30 seconds. Aspect ratios include adaptive, 16:9, 4:3, 1:1, 3:4 and 9:16.
Pricing is per second of output video: 20 gems/s at 480p, 40 gems/s at 720p and 80 gems/s at 1080p. A 5s 720p clip costs 200 gems; the cost display updates before you generate.
You can attach up to 10 reference images, 5 reference videos and 5 reference audios, plus 1 document file or 1 webpage link. Reference video duration plus output duration must not exceed 30 seconds, and audio must be paired with at least one image or video.
Upload a first frame and optionally a last frame image, and Wan 3.0 Prime generates a smooth transition video between them with precise motion control.
Both models accept exactly the same inputs and produce the same quality. Wan 3.0 Prime is the high-speed version with significantly shorter wait times, priced at 20/40/80 gems per second at 480p/720p/1080p versus 10/20/40 for the standard Wan 3.0.
