Skip to main content
Kandinsky 5.0 is a family of diffusion models for video and image generation developed by Kandinsky Lab. The Kandinsky 5.0 T2V Lite is a lightweight 2B parameter model that ranks among the top open-source video generation models, capable of generating videos up to 10 seconds long.
Make sure your ComfyUI is updated.Workflows in this guide can be found in the Workflow Templates. If you can’t find them in the template, your ComfyUI may be outdated.If nodes are missing when loading a workflow, possible reasons:
  1. You are not using the latest ComfyUI version (Nightly version)
  2. Some nodes failed to import at startup

​
Overview

Kandinsky 5.0 uses a latent diffusion pipeline with Flow Matching and features:
  • Diffusion Transformer (DiT): Main generative backbone with cross-attention to text embeddings
  • Qwen2.5-VL and CLIP: Provides high-quality text embeddings
  • HunyuanVideo 3D VAE: Encodes and decodes video into a latent space
The model family includes multiple variants optimized for different use cases:
  • SFT model: Highest generation quality
  • CFG-distilled: 2× faster inference
  • Diffusion-distilled: 6× faster with minimal quality loss (16 steps)
  • Pretrain model: Designed for fine-tuning
All models are available in 5-second and 10-second video generation versions.

​
Model variants

​
Text-to-Video workflow

​
Kandinsky 5.0 Video Lite Text to Video

A lightweight 2B model that generates videos from English and Russian prompts with high visual quality. Kandinsky 5.0 Video Lite Text to Video workflow preview

Run on Comfy Cloud

Open in Comfy Cloud

Download Workflow

Download JSON or search “Kandinsky 5.0 Video Lite Text to Video” in Template Library

​
2. Manually download models

Text Encoders

Text Encoder: qwen_2.5_vl_7b_fp8_scaled.safetensors

Qwen2.5-VL 7B text encoder (FP8). Place in ComfyUI/models/text_encoders/

Text Encoder: clip_l.safetensors

CLIP-L text encoder. Place in ComfyUI/models/text_encoders/
Diffusion Model

Diffusion Model: kandinsky5lite_t2v_sft_5s.safetensors

Kandinsky 5.0 T2V Lite SFT diffusion model (5s). Place in ComfyUI/models/diffusion_models/
VAE

VAE: hunyuan_video_vae_bf16.safetensors

HunyuanVideo 3D VAE. Place in ComfyUI/models/vae/

​
Image-to-Video workflow

​
Kandinsky 5.0 Video Lite Image to Video

A lightweight 2B model that generates videos from English and Russian prompts with high visual quality. Kandinsky 5.0 Video Lite Image to Video workflow preview

Run on Comfy Cloud

Open in Comfy Cloud

Download Workflow

Download JSON or search “Kandinsky 5.0 Video Lite Image to Video” in Template Library
Input materials Upload this file to the matching LoadImage node:

crystal_flower.png

LoadImage node 11 · crystal_flower.png
crystal_flower.png

​
2. Manually download models

Text Encoders

Text Encoder: qwen_2.5_vl_7b_fp8_scaled.safetensors

Qwen2.5-VL 7B text encoder (FP8). Place in ComfyUI/models/text_encoders/

Text Encoder: clip_l.safetensors

CLIP-L text encoder. Place in ComfyUI/models/text_encoders/
Diffusion Model

Diffusion Model: kandinsky5lite_i2v_5s.safetensors

Kandinsky 5.0 I2V Lite diffusion model (5s). Place in ComfyUI/models/diffusion_models/
VAE

VAE: hunyuan_video_vae_bf16.safetensors

HunyuanVideo 3D VAE. Place in ComfyUI/models/vae/

​
Resources