WAN AI Models - Advanced Image & Video Generation

WAN 2.6 represents the latest generation with enhanced multi-shot segmentation, audio input support, and improved prompt understanding for both Chinese and English. WAN 2.2 models like the A14B series excel at photorealistic image generation and editing, while WAN 2.2 Animate models specialize in motion transfer and character replacement. On JAI Portal, you can test both versions side-by-side to determine which best suits your specific project needs without committing to separate subscriptions.

WAN models range from 0.06 credits (WAN 2.6 Text to Image) to 0.50 credits (WAN 2.6 Reference-to-Video) per generation. Fast versions like WAN 2.2 T2V Fast cost only 0.08 credits, while premium features like WAN 2.6 Image-to-Video cost 0.45 credits. You only pay for what you generate—no monthly subscriptions. New users receive 10 free starter credits to test any WAN model without a credit card.

Yes, WAN 2.6 Image to Video Spicy and WAN 2.6 Image-to-Video both support optional background audio input, allowing you to synchronize your generated videos with music or sound effects. This feature is particularly useful for creating social media content, advertisements, or presentations where audio-visual synchronization enhances the viewing experience. Simply upload your audio file along with your image and prompt when using these models on JAI Portal.

WAN 2.6 Reference-to-Video maintains subject consistency across video scenes by using 1-3 reference videos as visual guides. This ensures that characters, objects, or brand elements remain visually coherent throughout your generated content, making it ideal for creating video series, brand campaigns, or narrative content where consistency is crucial. This model costs 0.50 credits per generation and is available exclusively in the WAN 2.6 suite on JAI Portal.

WAN 2.2 Animate Move transfers motion and expressions from a reference video to your static images, creating animated versions that preserve your original character's appearance while adding dynamic movement. This is perfect for bringing illustrations, portraits, or character designs to life with realistic motion. The model costs 0.30 credits per generation and works by analyzing the motion patterns in your reference video and applying them to your input image.

Absolutely. Alibaba WAN 2.6 Image Edit specializes in precise photo editing based on text prompts, handling color correction, lighting adjustments, aesthetic restyling, background replacement, and object removal for just 0.06 credits. WAN 2.2 A14B Image-to-Image offers high-resolution photorealistic editing with precise prompt understanding for 0.15 credits. Both models deliver professional-quality edits without requiring traditional photo editing software skills.

For speed and cost optimization, use WAN 2.2 T2V Fast (0.08 credits for text-to-video) or WAN 2.2 I2V Fast (0.10 credits for image-to-video). These models are specifically optimized for faster generation times and lower credit costs while maintaining good quality. They're ideal for rapid prototyping, draft previews, or high-volume content creation where you need to generate multiple variations quickly before investing in higher-quality renders with standard WAN 2.6 models.

No, all content generated with WAN models on JAI Portal is completely watermark-free. You own the commercial rights to everything you create, whether it's images or videos. This applies to all 16 WAN models available on the platform, from budget-friendly Fast versions to premium Reference-to-Video generations. There are no hidden restrictions or additional fees for commercial usage.

Yes, JAI Portal's unified interface makes it easy to chain WAN models together. For example, you can generate an initial image with WAN 2.6 Text to Image, then animate it with WAN 2.6 Image-to-Video, and finally refine the result with WAN 2.2 Video-to-Video. You can also compare outputs from different WAN versions side-by-side to optimize your workflow. Since you only pay per generation with credits, experimenting with different model combinations is cost-effective compared to managing multiple separate subscriptions.