Wan2.2-I2V-A14B-Diffusers

The model is an image-to-video (I2V) generative diffusion model designed for high-quality video synthesis at 480P and 720P resolutions. It incorporates a Mixture-of-Experts (MoE) architecture with two specialized experts (high-noise and low-noise) to enhance model capacity while maintaining computational efficiency. The model supports both prompt-based and prompt-free video generation.

Key features

• MoE Architecture: Separates denoising stages across experts to improve layout and detail refinement.
• Resolution Support: Generates videos at 480P and 720P with 24fps.
• Possibilities and methods of use: Inference is possible on a single GPU with 80 GB of VRAM (optimizations can be used: `offload_model`, `convert_model_dtype`).
• Promptless generation: Generation is possible solely based on the input image, with optional extension of the prompt via third-party APIs (e.g. DashScope).

---

The model is a component of the video generation pipeline, consisting of:

• Text encoder: ~5.6B parameters,
• Transformer: 2*~14B parameters,
• VAE: ~127M parameters,

Total: ~34.3B parameters

Announce Date: 28.07.2025
Parameters: 14B
Experts: 2
Activated at inference: 14B
Developer: Alibaba Wan Team
License: Apache 2.0

Use our pre-built public endpoints for free to test inference and explore Wan2.2-I2V-A14B-Diffusers capabilities. You can obtain an API access token on the token management page after registration and verification.

Rent your own physically dedicated instance with hourly or long-term monthly billing.

We recommend deploying private instances in the following scenarios:

• maximize endpoint performance,
• enable full context for long sequences,
• ensure top-tier security for data processing in an isolated, dedicated environment,
• use custom weights, such as fine-tuned models or LoRA adapters.

Contact our dedicated neural networks support team at nn@immers.cloud or send your request to the sales department at sale@immers.cloud.