Wan2.1: Text-to-Video and Image-to-Video Generation by Alibaba Cloud
Wan2.1 is Alibaba Cloud's first open-source video generation model. Available in 14B and 1.3B sizes, covering T2V and I2V tasks under Apache 2.0.
Wan2.1
VideoText-to-VideoImage-to-VideoOpen SourceFirst-generation Wan video model — Alibaba Cloud's open-source video generation model released February 2025. Apache 2.0 licensed. Available in 14B and 1.3B parameter variants for text-to-video and image-to-video. ComfyUI native support. Also available as GGUF (city96) and Kijai WanVideoWrapper versions.
| Developer | Alibaba Cloud (Wan AI) |
| Release Date | 2025-02-28 |
| Architecture | Diffusion Transformer (DiT) |
| License | Apache-2.0 |
| Model Sizes | 14B (full) / 1.3B (lightweight) |
| Capabilities | Text-to-Video (14B/1.3B), Image-to-Video (14B) |
| Output Resolutions | 480P (I2V), 720P (T2V & I2V) |
Guides and workflows related to this model series.
Complete Wan2.1 Fun Control ComfyUI tutorial: I2V, T2V and V2V video generation with official, Kijai and GGUF workflows. Covers model setup and tuning.
This tutorial details how to use Wan2.1 Video LoRA models in ComfyUI
Download Wan2.1 ComfyUI workflow: wan2.1_i2v_480p_14B_fp16.safetensors, wan2.1_t2v_14B_fp16.safetensors, wan_2.1_vae.safetensors, GGUF and 5B variants. T2V, I2V and V2V guide.
Comments
Sign in with GitHub to join the discussion.