Wan models on AI Gateway

Alibaba's Wan video generation models are now available through Cloudflare's AI Gateway, supporting text-to-video, image-to-video, and reference-to-video generation.

• •Six model variants are available: wan-v2.6-t2v, i2v, i2v-flash, r2v, r2v-flash, and wan-v2.5-t2v-preview
• •Character Reference (R2V) allows extracting appearance and voice from reference images/videos to generate new scenes
• •Flash variants offer faster generation; supported resolutions are 480p, 720p, and 1080p
• •Integration via AI SDK 6 using experimental_generateVideo with model, prompt, resolution, and duration parameters
• •No-code experimentation is available through the AI Gateway playground for comparing providers and tweaking prompts