Image Generation Models: A Detailed Guide
This platform presents various models for generating, editing, and stylizing images. Below you will find a detailed description of each model and their features.
🎨 Kandinsky 4.1 / 3.1 / 3.0 / 2.2
Purpose: Creating images from text with a focus on the Russian cultural code.
General:
All versions of Kandinsky support image generation from text and also can:
✏️ Inpainting – drawing inside the image.
🌄 Outpainting – extending the image beyond the original boundaries.
Differences between versions:
Recommended:
– Choose 4.1 for complex requests and modern trends.
– Use 2.2 for authorial painterly and graphic works.
Generation examples
🖼️ ControlNet: Image-to-Image
Purpose: Changing the style of the original image.
🎨 Preserves structure and pose
💬 Change the style through text (e.g. "in pixel art style")
Workflow:
Image (input image) → Image (output image)
+ Text (style)
Usage example: Upload a photo → Write “in anime style” → Get a result with the same poses and composition.
🌐 Flux
Purpose: Creating images from text with a focus on modern trends and creativity.
💡 Ideal for unconventional, futuristic, and trendy concepts
🧠 Well understands fashionable visual solutions
Excellent for creating complex, detailed textures in special effects and design.
Workflow:
Text → Image
Example prompts:
“3D avatar in metaverse style”
“Surreal city made of light”