Overview
Wan2.2-TI2V-5B is a lightweight text-and-image-to-video model. Give it a prompt plus a reference frame or style image and it generates coherent, temporally stable clips with preserved identity, controllable camera/motion, and fast iteration for cost-sensitive workflows.
Description
Wan2.2-TI2V-5B combines text guidance with a visual reference so you can steer both what happens and how it looks in the resulting video. Start with a prompt and a key image—character art, product shot, brand styleframe—and the model animates a short sequence that keeps identity, materials, and composition consistent while following your directions for motion, pacing, and camera moves. Compared with heavier Wan variants, the 5B configuration prioritizes responsiveness and budget: clips render quickly, edits are iterative rather than destructive, and you can extend a take, replace backgrounds, or inpaint/outpaint regions without losing continuity. It handles small details and typography reliably enough for social content, demos, and campaign variations, and exports cleanly into standard post workflows. Teams pick TI2V-5B when they need flexible, reference-faithful video generation with the speed and cost profile to scale.
About Alibaba
Chinese e-commerce and cloud leader behind Taobao, Tmall, and Alipay.
Website:
alibaba.com
Related Models
Last updated: September 23, 2025