Skip to main content

Unlock 10X Faster Image Generation with the Latent Consistency Model

Unlock 10X Faster Image Generation with the Latent Consistency Model banner banner

What is the Latent Consistency Model

The latent Consistency Model (LCM) is a type of distilled diffusion model. With LCM, you can generate high-fidelity images much faster than standard diffusion models in a very short inference time of 1 second to generate 10 images.

  • It's trained to predict the output of diffusion models based on their early iteration stages, instead of running the full diffusion process from start to finish. This allows extremely fast generation.
  • In tests, LCM has produced images comparable in quality to standard diffusion models, but in as little as 4 steps instead of hundreds. This represents a huge speedup.
  • It works by learning the behavior of diffusion models during their early iterations. It then uses this knowledge to consistently predict the final output without needing to run the full diffusion steps.
  • This demonstrates how model distillation techniques can accelerate diffusion-based generation when applied effectively.
  • LCM was developed by researcher Simian Luo and provides a proof-of-concept for balancing speed and quality in AI image generation.
  • Ongoing work aims to enhance LCM further, such as adding negative prompting for even greater control over image properties.

How can the Latent Consistency Model influence future image generation

Where other models may need dozens of steps to generate a comparable result, LCM can achieve this in just 4 steps, taking 1 second to generate 10 images! banner banner

This level of speed without sacrificing quality is almost unbelievable. The LCM demonstrates how powerful distillation techniques can be for accelerating AI generation when applied effectively. It's revolutionizing how we think about the relationship between speed and quality for diffusion models.

How to Use the Latent Consistency Model and API

🤗 Get started: Latent Consistency Model API for Text-to-imageopen in new window

🔥 Try LCM Demo: Visit Playgroundopen in new window for access to more models and features.

🔥 Thanks for @0xbitches. See the link: SD-Webuiopen in new window and ComfyUIopen in new window.

🔥 Check API Docs @novita.aiopen in new window . See the link: LCM APIopen in new window

Step-by-Step Guide for Using the API

Please set the Content-Type header to application/json in your HTTP request to indicate that you are sending JSON data. Currently, only JSON format is supported.


curl --location --request POST '' \
--header 'Authorization: Bearer {{key}}' \
--header 'Content-Type: application/json' \
--data-raw '{
   "prompt": "Glowing jellyfish floating through a foggy forest at twilight",
   "height": 512,
   "width": 512,
   "image_num": 2,
   "steps": 8,
   "guidance_scale": 7.5

HTTP status codes in the 2xx range indicate that the request has been successfully accepted, code 400 means requst params error, while status codes in the 5xx range indicate internal server errors.


    "images": [
            "image_file": "{{Base64 encoded image}}",
            "image_type": "jpeg"
            "image_file": "{{Base64 encoded image}}",
            "image_type": "jpeg"


In conclusion, the Latent Consistency Model is a distilled diffusion model trained to generate high-fidelity images exceptionally fast through consistency with standard diffusion model behavior during early iterations. It represents an exciting approach to rapid, high-quality AI image generation.

I hope you have a good experience. If you have any other questions, feel free to reach out to me on Discordopen in new window.

novita.aiopen in new window provides Stable Diffusion API and hundreds of fast and cheapest AI image generation APIs for 10,000 models.🎯 Fastest generation in just 2s, Pay-As-You-Go, a minimum of $0.0015 for each standard image, you can add your own models and avoid GPU maintenance. Free to share open-source extensions.