Skip to main content
Endpoint: POST https://fal.run/fal-ai/z-image/turbo/lora Endpoint ID: fal-ai/z-image/turbo/lora

Try it in the Playground

Run this model interactively with your own prompts.

Quick Start

Tongyi-MAI’s Z-Image Turbo with LoRA support delivers 6B-parameter text-to-image generation at $0.0085 per megapixel with 8-step inference. Apply up to 3 custom LoRA weights at inference time without retraining, enabling style consistency and brand adaptation at approximately 118 generations per dollar on fal. Built for: Custom style application | Brand-consistent generation | Character consistency workflows | Rapid design iteration

LoRA Flexibility Without Training Overhead

Z-Image Turbo LoRA adds inference-time style customization to the base model’s speed advantages. Apply pre-trained LoRA adapters directly through the API, combining multiple style influences in a single generation call without touching the underlying model weights. What this means for you:
  • Apply up to 3 LoRA weights simultaneously: Combine custom styles, character adapters, and brand guidelines in a single generation request through the loras parameter
  • Batch generation at scale: Generate up to 4 images per request with configurable inference steps (1-8 range), optimizing the speed-quality tradeoff for your use case
  • Acceleration options: Choose between “none”, “regular”, or “high” acceleration modes to balance generation speed against output fidelity
  • Production-ready safety: Built-in safety checker (enabled by default) filters NSFW content automatically, with optional prompt expansion for enhanced detail at +$0.0025 per request
  • Flexible output formats: Generate images in JPEG, PNG, or WebP with landscape (4:3), portrait (3:4), or square (1:1) aspect ratios

Performance That Scales

Z-Image Turbo LoRA adds minimal overhead to base model pricing while enabling custom style application at inference time.

Technical Specifications

API Documentation | Quickstart Guide

How It Stacks Up

Z-Image Turbo – The base Z-Image Turbo endpoint runs at 0.005/MPformaximumcostefficiencywhenstylecustomizationisntrequired.TheLoRAvariantadds0.005/MP for maximum cost efficiency when style customization isn't required. The LoRA variant adds 0.0035/MP overhead to enable custom style application, character consistency, and brand adaptation at inference time. AuraFlow – Z-Image Turbo LoRA prioritizes cost efficiency at $0.0085/MP with 8-step inference and runtime style customization. AuraFlow emphasizes open-source flexibility and longer inference paths for applications requiring maximum creative control and community-driven development. FLUX.2 [dev] LoRA – Z-Image Turbo LoRA delivers comparable style customization at 0.0085/MPversusFLUX.2[dev]LoRAs0.0085/MP versus FLUX.2 [dev] LoRA's 0.021/MP, making it 2.5x more cost-efficient for high-volume workflows. FLUX.2 [dev] LoRA offers higher resolution outputs and more sophisticated prompt interpretation for applications where output quality justifies the premium.

Capabilities

  • Text prompt input
  • Configurable resolution
  • Adjustable inference steps
  • Reproducible generation (seed)
  • Synchronous mode
  • Batch generation
  • Safety checker
  • LoRA support

API Reference

Input Schema

string
required
The prompt to generate an image from.
ImageSize | Enum
default:"landscape_4_3"
The size of the generated image. Default value: landscape_4_3Possible values: square_hd, square, portrait_4_3, portrait_16_9, landscape_4_3, landscape_16_9
integer
default:"8"
The number of inference steps to perform. Default value: 8Range: 1 to 8
integer
The same seed and the same prompt given to the same version of the model will output the same image every time.
boolean
default:"false"
If True, the media will be returned as a data URI and the output data won’t be available in the request history.
integer
default:"1"
The number of images to generate. Default value: 1Range: 1 to 4
boolean
default:"true"
If set to true, the safety checker will be enabled. Default value: true
OutputFormatEnum
default:"png"
The format of the generated image. Default value: "png"Possible values: jpeg, png, webp
AccelerationEnum
default:"regular"
The acceleration level to use. Default value: "regular"Possible values: none, regular, high
boolean
default:"false"
Whether to enable prompt expansion. Note: this will increase the price by 0.0025 credits per request.
list<LoRAInput>
default:""
List of LoRA weights to apply (maximum 3).

Output Schema

list<ImageFile>
required
The generated image files info.
Timings
required
The timings of the generation process.
integer
required
Seed of the generated Image. It will be the same value of the one passed in the input or the randomly generated that was used in case none was passed.
list<boolean>
required
Whether the generated images contain NSFW concepts.
string
required
The prompt used for generating the image.

Input Example

Output Example

Limitations

  • image_size restricted to: square_hd, square, portrait_4_3, portrait_16_9, landscape_4_3, landscape_16_9
  • num_inference_steps range: 1 to 8
  • num_images range: 1 to 4
  • output_format restricted to: jpeg, png, webp
  • acceleration restricted to: none, regular, high
  • Content moderation via safety checker