ZImageApp v2.1.0 · Apple Silicon · macOS 15+

Qwen Image 2.1 Turbo on Mac with MLX

Use the dedicated Qwen Image 2.1 Turbo model locally on Mac with native Swift / MLX inference.

Developer ID signed and Apple notarized. Models download separately.

Qwen Image 2.1 Turbo on your Mac

Qwen-Image-2.1-Turbo is a separate model option in ZImageApp for Apple Silicon. Select its compatible q4 MLX package to generate images with the Turbo workflow, without setting up a Python inference environment.

Choose the Turbo entry directly rather than treating every acceleration LoRA as the Turbo model. CacheDiT remains off by default for Qwen and can be enabled when needed. Generation speed and memory use depend on resolution, model settings and your Mac.

How to get started

Download and install the Apple Silicon DMG.
Select Qwen Image 2.1 Turbo and download or choose its compatible model package.
Use the Turbo model settings, enter a prompt and generate locally.

Is Qwen Image 2.1 Turbo the same as adding a LoRA?

The app offers Qwen Image 2.1 Turbo as its own model choice. Use its matching model package; acceleration LoRAs for the standard model are a separate workflow.