Wan 2.2 Series Now Available on SiliconFlow：Generate Stable, Realistic and Cinematic Videos

Aug 16, 2025

SiliconFlow is excited to bring the Wan 2.2 series to our platform — a major upgrade to Wan's visual generative models, featuring significantly improved performance and superior visual quality. We provide two models from this series: Wan2.2-T2V-A14B and Wan2.2-I2V-A14B. Designed to power the next wave of creative AI applications from concept art to commercial visuals, these models enabling users to generate more detailed, realistic, and imaginative content than ever before.

In addition to Wan 2.2, SiliconFlow also supports the previously released Wan 2.1 series, giving users the flexibility to select the most suitable model for their specific needs.

With SiliconFlow's Wan 2.2 API, you can expect:

Cost-Effective Pricing: Wan2.2-T2V-A14B $0.29 / video;Wan2.2-I2V-A14B $0.29 / video.
Resolution Support: Wan2.2-T2V-A14B: 480P & 720P (5s);Wan2.2-I2V-A14B: 480P & 720P.

Model Highlights in WAN 2.2 Series

SiliconFlow offers two powerful models from the WAN 2.2 series. You can select the appropriate model based on their specific task requirements and desired functionality:

T2V-A14B:
- Delivers high-quality video generation.
- Outperforms leading commercial models on Wan-Bench 2.0.

I2V-A14B:
- Produces more stable, coherent videos.
- Reduces unrealistic camera movements.Enhanced support for diverse stylized scenes.

These improvements in WAN 2.2 are further validated through benchmark performance. In comparative tests on WAN-Bench 2.0, WAN2.2-T2V-A14B outperforms leading closed-source commercial models such as Sora and Hailuo 02 across several critical dimensions.

These superior results stem from WAN 2.2's key technical innovations:

Effective MoE Architecture: WAN 2.2 introduces a Mixture-of-Experts (MoE) architecture into video diffusion models, separating the denoising process across timesteps with specialized expert models to enlarge overall model capacity while maintaining computational cost.
Cinematic-level Aesthetics: The model incorporates meticulously curated aesthetic data with detailed labels for lighting, composition, contrast, and color tone, enabling precise and controllable cinematic style generation.
Complex Motion Generation: Trained on significantly larger data with +65.6% more images and +83.2% more videos compared to WAN 2.1, enhancing generalization across motions, semantics, and aesthetics.

With Wan 2.2 series integrated into SiliconFlow, you can now benefit from these improvements in your own projects.

Real-world Performance On SiliconFlow

Cinematic Vision Control

prompt: A purely visual and atmospheric video piece focusing on the interplay of light and shadow, with a corn train as the central motif. Imagine a stage bathed in dramatic, warm spotlights, where a corn train, rendered as a stark silhouette, moves slowly across space. The video explores the dynamic interplay of light and shadow cast by the train, creating abstract patterns, shapes, and illusions that dance across the stage. The soundtrack should be ambient and minimalist, enhancing the atmospheric and abstract nature of the piece.

Sweeping Motion

prompt: A dynamic group of diverse hip-hop dancers explodes across a vast stage bathed in vibrant neon lights, casting side-lit halos around their silhouettes. Wide cinematic shot captures synchronized movements, vibrant energy, and youthful expressions. Fast-paced camera work mirrors the beat, showcasing intricate footwork, explosive power, and the unity of the ensemble.

Precise Prompt Following

prompt: Watercolor style, the wet suminagashi inks slowly spread into the shape of an island on paper, with the edges continuously blending into delicate textural variations. A tiny paper boat floats in the direction of the water flows towards the still-wet areas, creating subtle ripples around it. Centered composition with soft natural light pouring in from the side, revealing subtle color gradations and a sense of movement.

Get Started Immediately

Explore: Try Wan 2.2 series in the SiliconFlow playground.
Integrate: Use our OpenAI-compatible API. Explore the full API specifications in the SiliconFlow API documentation.

import requests

url = "https://api.siliconflow.com/v1/video/submit"

payload = {
    "model": "Wan-AI/Wan2.2-T2V-A14B",
    "prompt": "an island near sea, with seagulls, moon shining over the sea, light house, boats int he background, fish flying over the sea",
    "image_size": "960x960"
}
headers = {
    "Authorization": "Bearer <token>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.json())