Wan 2.2
Wan 2.2 is a video generation model from Wan that converts source images into videos with customizable LoRA fine-tuning.
Wan
WaveSpeedImage-to-video generation with LoRA support
Wan 2.2 is a video generation model published by Wan and served through the WaveSpeed provider on MindStudio. Released in July 2025, it accepts a source image as input and generates video output, with controls for resolution, duration, seed, and negative prompting. Pricing ranges from $0.20 to $0.64 per video depending on output settings, and the model supports a context window of 1,000 tokens.
A notable feature of Wan 2.2 is its support for LoRA (Low-Rank Adaptation) fine-tuning, which allows users to apply custom style or subject adaptations to the generated video. This makes it well-suited for workflows that require consistent visual styles or character appearances across multiple generations. The model is best used for image-to-video animation tasks where reproducibility and stylistic control are priorities.
What Wan 2.2 supports
Image-to-Video
Animates a source image into a video clip. Accepts an image URL as the primary input and generates motion from the still frame.
LoRA Fine-Tuning
Supports Low-Rank Adaptation (LoRA) weights to apply custom styles or subjects to generated videos. Multiple LoRAs can be specified per generation.
Resolution Control
Offers selectable output resolution options to match target display or quality requirements. Resolution is chosen via a dropdown input at generation time.
Duration Selection
Allows users to choose the length of the generated video clip via a select input. Duration choice affects the final per-video cost within the $0.20–$0.64 range.
Negative Prompting
Accepts a negative prompt text field to suppress unwanted visual elements or styles from the output. Helps steer generation away from specific artifacts or content.
Seed Control
Supports a seed input for reproducible video generation results. Using the same seed with identical settings produces consistent outputs.
Ready to build with Wan 2.2?
Get Started FreeCommon questions about Wan 2.2
How much does it cost to generate a video with Wan 2.2?
Wan 2.2 is priced between $0.20 and $0.64 per video. The exact cost depends on the resolution and duration settings selected at generation time.
What inputs does Wan 2.2 require?
The model accepts a source image URL, resolution selection, duration selection, optional LoRA weights, an optional negative prompt, and an optional seed value.
Does Wan 2.2 support LoRA customization?
Yes. Wan 2.2 includes a dedicated LoRA input, allowing users to apply Low-Rank Adaptation weights to customize the style or subject of generated videos.
What is the context window for Wan 2.2?
Wan 2.2 has a context window of 1,000 tokens.
When was Wan 2.2 released?
Wan 2.2 was released in July 2025 and is currently available on MindStudio via the WaveSpeed provider.
What people think about Wan 2.2
Community reception to Wan 2.2 on Reddit has been notably positive, with users highlighting the model's motion quality and its compatibility with third-party tools like SVI 2.0 Pro and Time-to-Move. The most upvoted thread, with over 6,000 upvotes, showcases motion transfer results, while a thread about seamless long-form video generation at 1280x720 over 20 seconds attracted over 2,100 upvotes and nearly 400 comments.
Users frequently discuss Wan 2.2 in the context of open-source workflows and ComfyUI pipelines, with particular interest in the Wan-Animate module for single-image character animation. No significant technical limitations or quality concerns dominate the threads, though discussions reflect that generation times and hardware requirements remain relevant considerations for longer or higher-resolution outputs.
Z-Image with Wan 2.2 Animate is my wet dream
SVI 2.0 Pro for Wan 2.2 is amazing, allowing infinite length videos with no visible transitions. This took only 340 seconds to generate, 1280x720 continuous 20 seconds long video, fully open source. Someone tell James Cameron he can get Avatar 4 done sooner and cheaper.
Time-to-Move + Wan 2.2 Test
Parameters & options
Up to 3 LoRAs.
Description of what to exclude from the video.
A specific value that is used to guide the 'randomness' of the generation.
Explore similar models
WanWan 2.7
Wan 2.7 is a video generation model from Alibaba that supports text-to-video, image-to-video, video editing, and audio generation.
WanWan 2.6
Wan 2.6 is a video generation model that creates videos from source images and audio with multi-shot control.
WanWan 2.5
Wan 2.5 is a video generation model that animates source images with optional audio input, priced at $0.05–$0.15 per second.
LightricksLTX-2 19b
LTX-2 19b is a 19-billion-parameter video generation model from Lightricks that supports image-to-video synthesis and LoRA fine-tuning.
X.aiGrok Imagine 1.5
Grok Imagine 1.5 is a video generation model from X.ai that accepts text, image, and video inputs to produce generated video output.
AlibabaHappyHorse 1.0
HappyHorse 1.0 is a video generation model from Alibaba that supports image-to-video, video editing, and reference image inputs.
Start building with Wan 2.2
No API keys required. Create AI-powered workflows with Wan 2.2 in minutes — free.