Z-Image TurboGenerate Images at the Speed of Thought
6B Parameters. Sub-Second Speed.
Images in Under a Second
Z-Image Turbo compresses inference to just 1-8 steps, generating images far faster than standard diffusion models. Configurable step count lets you trade quality for speed based on your use case: 1-step for prototyping, 8-step for final assets.
Precise Prompt Adherence
From photorealistic portraits to abstract art, Z-Image Turbo handles a wide range of styles and compositions. The 6B-parameter architecture delivers strong prompt following, so what you describe is what you get, without excessive prompt engineering.
Beyond Text-to-Image
Z-Image Turbo supports image-to-image transformation, inpainting, ControlNet (edge, depth, pose), and custom LoRA fine-tuning. Chain these together in production workflows for style transfer, product variants, or iterative refinement at scale.
One model, many modes
Z-Image offers specialized endpoints for every image generation workflow. Pick the one that fits your use case.
See what Z-Image Turbo can create
Copy any prompt below and try it yourself in the playground.
"A hyper-realistic close-up portrait of a tribal elder, intricate white chalk patterns, headdress of dried flowers and seed pods, warm firelight reflected in dark soulful eyes, Kodak Portra 400 film grain"
"Isometric view of a tiny Japanese zen garden on a floating island, miniature bonsai trees, raked sand patterns, koi pond with golden fish, soft morning light, tilt-shift effect"
"An astronaut standing on the surface of Mars, Earth visible in the salmon-pink sky, dramatic volumetric lighting, cinematic composition, ultra-wide angle"
"Watercolor painting of a Venetian canal at sunset, gondolas reflecting in shimmering water, warm amber and rose tones, loose impressionistic brushstrokes"
A few lines of code.
Instant images.
fal.ai handles the infrastructure: fast inference, auto-scaling, and a developer-friendly API. No GPUs to manage.
- Serverless: scales to zero, scales to millions
- Pay per megapixel, no minimums
- Python and JavaScript SDKs, plus REST API
import fal_client
result = fal_client.run(
"fal-ai/z-image/turbo",
arguments={
"prompt": "A serene mountain lake
at golden hour, photorealistic",
}
)
# result.images[0].url → your generated imageCommon questions about Z-Image Turbo
What can I create with Z-Image Turbo?
Z-Image Turbo supports text-to-image generation, image-to-image transformation, inpainting (editing parts of an image with a mask), and ControlNet-guided generation using edge, depth, or pose inputs. You can also apply custom LoRA weights to any of these modes for style or subject fine-tuning. Output resolutions go up to 4 megapixels with flexible aspect ratios.
What is the difference between Z-Image Turbo and Z-Image Base?
Z-Image Turbo is optimized for speed, using a compressed 1-8 step inference pipeline at $0.005 per megapixel. It is ideal for high-volume workflows, rapid prototyping, and production environments where throughput and cost matter most. Z-Image Base is the foundation model at $0.01 per megapixel, offering stronger generative diversity and broader stylistic coverage when maximum quality is the priority.
How much does Z-Image Turbo cost on fal.ai?
Pricing is pay-per-megapixel with no minimums or subscriptions. Text-to-image starts at $0.005 per megapixel (Turbo) or $0.01 per megapixel (Base). LoRA endpoints cost $0.0085/MP, ControlNet costs $0.0065/MP, and ControlNet + LoRA costs $0.01/MP. That means you get roughly 200 megapixels per $1.00 on Turbo. Optional prompt expansion adds $0.0025 per request.
Does Z-Image Turbo support LoRA fine-tuning?
Yes. You can apply up to 3 custom LoRA weights simultaneously on the text-to-image, image-to-image, and ControlNet endpoints. This is useful for consistent brand styles, character references, or specialized aesthetics. You can train your own LoRAs using the Z-Image Trainer on fal.ai.
What resolutions and formats does Z-Image Turbo support?
Z-Image Turbo generates images up to 4 megapixels with configurable aspect ratios from square to ultrawide. Output formats include JPEG, PNG, and WebP. You can also batch-generate up to 4 images per API call for faster iteration.
How do I get started with the API?
Install the fal.ai SDK (Python or JavaScript), grab an API key from your dashboard, and make your first request in three lines of code. The API is serverless, so there are no GPUs to manage and no infrastructure to set up. Check the API documentation for all available parameters.
Can I use Z-Image Turbo for commercial projects?
Yes. Images generated through the fal.ai API can be used in commercial projects. Check fal.ai's terms of service for full details on usage rights and licensing.
Ready to generate?
Start creating images with Z-Image Turbo on fal.ai.