text/image-to-video generation
Wan2.2 is an open-source, large-scale text/image-to-video generation model family from Wan-Video using a mixture-of-experts diffusion architecture with advanced aesthetic control.
Same problem, same approach. Swapping one for another is a config change, not a rewrite.
choose it when you want a recently updated, lighter-weight standalone video generation model from a different major vendor, accepting an unasserted license.
HunyuanVideo-1.5: A leading lightweight video generation model
choose it when you need a unified framework that can both generate and edit videos with an MLLM-based planner.
Bernini is a unified framework for video generation and editing that combines an MLLM-based semantic planner with a DiT-based renderer.
choose it when you need extremely long or infinite-length video outputs rather than standard short clips.
SkyReels-V2: Infinite-length Film Generative model
An earlier or more limited way to do the job. Still the right call when you need something small, proven, or CPU-only.
choose it when the task is subject-consistent generation across shots, narrower than general text-to-video.
Phantom: Subject-Consistent Video Generation via Cross-Modal Alignment
choose it for reference-driven human image animation rather than generic text-to-video.
UniAnimate-DiT: Human Image Animation with Large-Scale Video Diffusion Transformer
small research-stage video model with little maintenance; choose it only for lightweight experimentation, not production replacement.
MoviiGen 1.1: Towards Cinematic-Quality Video Generative Models
choose it when you need efficient pyramidal flow-matching generation and can accept a codebase unmaintained for over 18 months.
[ICLR 2025] Pyramidal Flow Matching for Efficient Video Generative Modeling
choose it when you need 3D-consistent character animation with pose control, a narrower task.
SCAIL: Towards Studio-Grade Character Animation via In-Context Learning of 3D-Consistent Pose Representations (CVPR 2026 Findings)
choose it for portrait animation and expression retargeting instead of general video synthesis.
Bring portraits to life!
Does the same job, but ships as an app. Useful to a person, not swappable into a codebase.
choose it when you want a packaged automated short-video engine instead of integrating a model into your own codebase.
๐ AI ๅ จ่ชๅจ็ญ่ง้ขๅผๆ | AI Fully Automated Short Video Engine
These projects were analysed and named Wan2.2 among their alternatives. The relationship is not symmetric โ how Wan2.2 rates them is a separate judgement, made when Wan2.2 is analysed in its own right.
calls Wan2.2 โPick it for a strong Apache-2.0 T2V/I2V family with a wide range of model sizes to trade off quality and VRAM.โ
text and image to video generation: CogVideoX (2024) and CogVideo (ICLR 2023)
calls Wan2.2 โpick it when you want an Apache-2.0 text-to-video model with broad ecosystem support and active releasesโ
HunyuanVideo: A Systematic Framework For Large Video Generation Model
calls Wan2.2 โChoose for a state-of-the-art large-scale video generation model when you don't need MMagic's unified zoo.โ
OpenMMLab Multimodal Advanced, Generative, and Intelligent Creation Toolbox. Unlock the magic ๐ช: Generative-AI (AIGC), easy-to-use APIs, awsome model zoo, diffusion models, for text-to-image generation, image/video restoration/enhancement, etc.