Skip to content
BackComfyUI Model Pack Detail
Model Info
TypeModel Pack
LicenseApache-2.0
Files13
RepositoryHuggingFace ↗

Wan2.2-S2V-14B-GGUF

Note: All model files below must be downloaded via the LM Downloader client. If you do not have the client, Download >>>

Wan2.2-S2V-14B is an audio-driven cinematic video generation model launched by Wan-AI in August 2025, upgraded from the Wan2.2 series. It adopts a Mixture-of-Experts (MoE) architecture, enabling it to generate 480P and 720P resolution videos by combining audio, reference images, and optional text prompts. It also supports pose-driven generation to match specific motion sequences. Note: It is recommended to choose a model file of appropriate size based on your needs and VRAM capacity. We tested with an NVIDIA RTX 5060ti (16GB VRAM), and the quality generated by Wan2.2-S2V-14B-Q4_K_M.gguf was quite good.

NameSizeTarget Dir
Wan2.2-S2V-14B-Q2_K.gguf9.51 GBunet
Wan2.2-S2V-14B-Q3_K_S.gguf10.72 GBunet
Wan2.2-S2V-14B-Q3_K_M.gguf11.39 GBunet
Wan2.2-S2V-14B-Q4_0.gguf12.77 GBunet
Wan2.2-S2V-14B-Q4_K_S.gguf12.96 GBunet
Wan2.2-S2V-14B-Q4_1.gguf13.47 GBunet
Wan2.2-S2V-14B-Q4_K_M.gguf13.86 GBunet
Wan2.2-S2V-14B-Q5_K_S.gguf14.35 GBunet
Wan2.2-S2V-14B-Q5_0.gguf14.52 GBunet
Wan2.2-S2V-14B-Q5_K_M.gguf15.00 GBunet
Wan2.2-S2V-14B-Q5_1.gguf15.23 GBunet
Wan2.2-S2V-14B-Q6_K.gguf16.21 GBunet
Wan2.2-S2V-14B-Q8_0.gguf19.62 GBunet