Z-Image Turbo ControlNet: Precision Control for Image-to-Image

Z-image turbo is an open-source model and it's great because it is flexible, results are high fidelity, provides highly detailed images pre-upscale, and is super lightweight. Controlnet is great for crafting complex compositions & allowing the user to create images with great precision.

Key Inputs

Besides resolution and number of images per batch (batch_size), these are the key inputs for controlling the output:

• Positive Prompt: Write a detailed prompt describing your desired output.

• Negative Prompt: Jot down things you don't want in your image.

• Optional Image Input: Toggle this on to enable image-to-image and upload an image you’d like to transform.

• Denoise: Controls how much the output differs from the original image when using image-to-image.

  • High values (e.g., 0.9) → More different from the original

  • Low values (e.g., 0.1) → Closer to the original

• steps: Recommended default value of 8 gets all the quality you need for max speed.

Z-Image Workflows Collection

Z-Image isn't just a standalone model - it is becoming a complete production suite for anything we need for image generation and editing. While the base model is already changing the game right now, the roadmap for the next few weeks will cement Z-Image as the undisputed king for all things image generation:

1. Z-Image Turbo - collection

2. Z-Image Turbo - Text2Image: Available now

3. Z-Image Turbo - Image2Image: Available now

4. Z-Image & Dype for 4k resolution & high fidelity images - Available Now

5. Z-Image Turbo Upscale: Coming soon

6. Z-Image Turbo - LoRA Training: Coming soon here on Floyo before end of Dec 2025.. We will have a full tutorial on Floyo for training consistent characters, objects, and styles!

We don't need anything else for image generation and editing; Z-Image has it all. It is becoming a key piece of every production workflow we run.