Everything you need
to launch your AI product

Customize, deploy, and scale models on serverless GPUs with the world's first Python Cloud. Improve your customer experience with lightning fast cold starts and pay only for what you use.

import fal

@fal.function(
    requirements=["diffusers[torch]", "transformers"],
    machine_type="GPU",
    keep_alive=60,
)
def run_stable_diffusion(options: DiffusionOptions) -> Result:
		# Load the Diffusers pipeline
    pipe = load_model()

		# Perform the inference process
    result = pipe(options.prompt, num_inference_steps=options.steps)

		# Upload the image and return it
    image = Image.from_pil(result.images[0])
		return Result(image=image)

run_stable_diffusion("An astronaut riding a horse")
Normalstable_diffusion.py
$ python stable_diffusion.py
18:28:17 Generating new image

Pricing

Ship custom model endpoints with fine-grained control over idle timeout, max concurrency and autoscaling.

  • CPU: 10
  • Memory: 64GB
  • GPU: A100 (40GB VRAM)
  • Total Price: $0.00195/s

Unit Price

CPU
$0.00005/s
Memory
$0.000007/s
GPU A100
$0.001/s
GPU T4
$0.0002/s
GPU L4
$0.0003/s
Storage
$2/GB/month

Join our community

Join the discussion around our product and help shape the future of AI.