Today marks a significant milestone in our journey to redefine generative video. We are thrilled to announce the release of Wan 2.2 , the latest iteration of our core generation engine.
The primary focus of this update has been performance. Our engineering team has implemented architectural improvements that allow for 2x faster video generation speeds. This means less time waiting and more time creating, enabling a smoother and more efficient workflow for creators everywhere.
Overview of Wan 2.2
Wan 2.2 is an advanced video generation model released in mid-2025 that pushes the boundaries of open-source video AI. It is part of a comprehensive suite that includes models for text-to-video (T2V) and image-to-video (I2V) tasks.
- Core Goal: To make high-end video production faster, cheaper, and more accessible.
- Key Improvement: It features a "fast" variant specifically optimized for quick turnaround times, often utilizing acceleration techniques like PrunaAI optimization.
2x Faster Generation
The primary breakthrough of Wan 2.2 is its revolutionary approach to inference speed.
- Wait Times: Users will experience up to a 50% decrease in rendering times for standard high-definition clips.
- Real-Time Iteration: This speed allows creators to experiment with different prompts and styles more rapidly, perfecting their vision in half the time.
Improved Motion and Character Consistency
Speed is nothing without quality. Wan 2.2 isn't just faster; it’s smarter.
- Fluid Physics: The model features improved temporal consistency, meaning hair, clothing, and background elements move with more natural physics.
- Synchronization: Working alongside our Character-3 technology, Wan 2.2 ensures that lip-syncing and facial expressions are perfectly timed with your audio tracks.
What's New in 2.2?
- Optimized Latent Diffusion: Enhanced processing layers that reduce computational overhead.
- Improved Frame Consistency: Better temporal stability across long sequences.
- Enhanced Detail Transfer: Sharper textures and more realistic lighting in complex scenes.
How to Use Wan 2.2 Today
We have made the transition seamless for all our creators..
- Automatic Integration: The Wan 2.2 engine is now the default model for all new projects in the Hedra Studio.
- Compatible with Elements: You can use Wan 2.2 with our pre-set Hedra Elements to quickly place high-quality characters into cinematic environments.
Ready to try Hedra?
Transform your ideas into cinematic video in seconds.