HunyuanVideo

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

HunyuanVideo is a large-scale open-source video generation model by Tencent, with 13 billion parameters. It features physical simulation and high text semantic fidelity, capable of generating videos with movie-level quality.

Hunyuan Video I2V >13 files

Name	Size	Target Dir
hunyuan-video-i2v-720p-Q3_K_S.gguf	6.09 GB	unet
hunyuan-video-i2v-720p-Q3_K_M.gguf	6.24 GB	unet
hunyuan-video-i2v-720p-Q4_0.gguf	7.74 GB	unet
hunyuan-video-i2v-720p-Q4_K_S.gguf	7.74 GB	unet
hunyuan-video-i2v-720p-Q4_K_M.gguf	7.88 GB	unet
hunyuan-video-i2v-720p-Q4_1.gguf	8.52 GB	unet
hunyuan-video-i2v-720p-Q5_0.gguf	9.30 GB	unet
hunyuan-video-i2v-720p-Q5_K_S.gguf	9.30 GB	unet
hunyuan-video-i2v-720p-Q5_K_M.gguf	9.45 GB	unet
hunyuan-video-i2v-720p-Q5_1.gguf	10.08 GB	unet
hunyuan-video-i2v-720p-Q6_K.gguf	10.95 GB	unet
hunyuan-video-i2v-720p-Q8_0.gguf	13.97 GB	unet
hunyuan-video-i2v-720p-BF16.gguf	25.65 GB	unet

Hunyuan Video VAE >2 files

Name	Size	Target Dir
hunyuan_video_vae_fp8_e4m3fn.safetensors	247 MB	vae
hunyuan_video_vae_bf16.safetensors	493 MB	vae

Flux Text Encoders >4 files

Name	Size	Target Dir
clip_l.safetensors	246 MB	text_encoders
t5xxl_fp8_e4m3fn.safetensors	4.89 GB	text_encoders
t5xxl_fp8_e4m3fn_scaled.safetensors	5.16 GB	text_encoders
t5xxl_fp16.safetensors	9.79 GB	text_encoders

llava llama3 Encoders >2 files

Name	Size	Target Dir
llava_llama3_fp8_scaled.safetensors	9.09 GB	text_encoders
llava_llama3_fp16.safetensors	16.1 GB	text_encoders

llava_llama3_vision >1 files

Name	Size	Target Dir
llava_llama3_vision.safetensors	649 MB	clip_vision