Get a weekly rundown of the latest AI models and research... subscribe!

Illusion Diffusion Hq


AI model preview image
The illusion-diffusion-hq is a model created by Monster Labs that uses the QrCode ControlNet on top of SD Realistic Vision v5.1. It operates by taking an input seed, image URL, dimensions (width, height and border), a prompt for image creation, a negative prompt to avoid, QR code content, QR code background color, number of inference steps and controlnet conditioning scale. The model generates an image as output, based on the prompts and guidance provided, by following the link generated under the output schema. This data can be used for creating AI-generated visual content like scenes, objects, or scenarios, and it emphasizes on generating detailed outputs with superior quality. For example, in the given schema, it will create an image of a detailed, high-quality medieval village scene with busy streets and a castle in the distance.

Use cases

The illusion-diffusion-hq AI model from Monster Labs is designed to generate high-quality, detailed visuals based on provided input. This AI model leverages QrCode ControlNet technology over an SD Realistic Vision model and uses a specific seed along with characteristics defined in the input to produce a specific image output. A key use case for this model is in graphic design, where it can be utilized to create unique and visually compelling scenes, given a certain seed and prompt. For instance, based on the provided schema, this model can generate an intricately detailed medieval village scene with bustling streets and a castle in the distance. Being able to render imagery with set prompts and seeds means it could be used for a range of digital art, design and marketing endeavours. From video game design, where it can flesh out detailed concept art, to marketing agencies needing custom graphics for their campaigns and product designs or even interior design studios seeking unique wall art. Additionally, it can be useful in educational scenarios, creating detailed historic or scientifically accurate scenes to aid understanding in a visually compelling way. Lastly, it could be used to create digital components for various products like puzzles, toys, or web-based interactive applications requiring custom visuals. While the model currently generates static images, with further refinement and development, it could potentially be adapted for dynamic visual generation, paving the way for applications in animated films or dynamic web content. Its negative prompts feature, which seemingly sets parameters for unwanted characteristics, may also find utility in applications where tailoring the visual output to meet very specific quality or aesthetic requirements is essential.



Cost per run
Avg run time

Creator Models

Realvisxl2 Lora Inference$?1,987
Wizardcoder 15b V1$?459
Vicuna 13b V1.3$?3,554
Wizardcoder Python 34b V1.0$?830

Similar Models

Try it!

You can use this area to play around with demo applications that incorporate the Illusion Diffusion Hq model. These demos are maintained and hosted externally by third-party creators. If you see an error, message me on Twitter.

Currently, there are no demos available for this model.


Summary of this model and related resources.

Model NameIllusion Diffusion Hq
Monster Labs QrCode ControlNet on top of SD Realistic Vision v5.1
Model LinkView on Replicate
API SpecView on Replicate
Github LinkView on Github
Paper LinkNo paper link provided


How popular is this model, by number of runs? How popular is the creator, by the sum of all their runs?

Model Rank
Creator Rank


How much does it cost to run this model? How long, on average, does it take to complete a run?

Cost per Run$-
Prediction Hardware-
Average Completion Time-