You can get this model, along with many others, by joining The Tinkerer on Whop. It’s one monthly membership, and you’ll also get early releases, private tools, and a bunch of extra stuff.
CyberIllustrious blends the artistic flair of Illustrious-XL with a big dose of CyberRealistic’s style. You still get all those cinematic, dreamy strokes, but now they come through with crisp, realistic detail.
If you want that painterly, elegant look but still want your images to feel sharp and lifelike, this is your model. It’s great for editorial vibes, soft glam, or even fantasy art that looks like a real photo. Basically, it just nails it.
It’s powerful, easy to use, and honestly? Prompting it is seriously fun.
📘 Get the most out of this model with the Illustrious / NoobAI Prompting: Master Guide.
⚙️ Recommended Settings
Sampling method: DPM++ SDE Karras / DPM++ 2M Karras - SGM Uniform / Euler a
VAE: is already Baked In
Sampling steps: 30+ Steps
Resolution: 896x1152 / 832x1216
CFG: 4-5About (DMD2)
The special DMD2 edition of CyberIllustrious is created for faster, cleaner generations while requiring fewer steps.
DMD2 takes the balanced improvements of the standard release and optimizes them for speed, precision, and responsiveness, making it ideal for both SFW and NSFW workflows.
Expect tighter results, smoother control, and a model that reacts beautifully to the classic Illustrious prompting style.
Recommended Settings
Sampler: LCM
Schedule Type: Exponential
Sampling Steps: 8 (12 for extra fine detail)
CFG Scale: 1.5
⚡ These settings are optimized for DMD2 and are different from my standard Illustrious workflow!
🧠 Prompting Made Easy
Positive Prompt:
masterpiece, best quality, high quality, ultra-detailed, realistic, ... rest of promptNegative Prompt examples (optional):
cartoon, illustration, anime, painting, CGI, 3D render, low quality, watermark, logo, labeldeformed, deformed face, low quality, bad quality, worst quality, (drawn, furry, illustration, cartoon, anime, comic:1.5), 3d, cgi, extra fingers, deformed teeth, watermark, logo🩺 Optional ADetailer Settings
Adetailer model: face_yolov9c.pt
If you only want the main face being refined set 'Mask only the top k largest' to 1.🔗 Resources
Backup location: HuggingFace
☕ Support the Madness
Liking what you see? Grab me a digital espresso (or a couple!). It keeps the creative energy going and helps bring more cool prompts and models your way.
⚠️ Heads Up
This model can spit out some mature stuff. Use it wisely, and please, no cursed creations—we’ve really seen enough of those.
💬 Want Better Prompts?
Use this custom ChatGPT to auto-generate polished prompts for CyberIllustrious:
🔗 [Try it now on ChatGPT]
Description
I've just made further refinements to this model.
FAQ
Comments (35)
It looks like the 3.5 download is bad. 3.0 works OK, but 3.5 won't generate images.
I may be having same problem. Are you using AMD GPU?
Seems to be working fine for me. Running ComfyUI on an AMD Ryzen 7 5800X3D and a Radeon RX 6800 XT.
Duplicate post, but I want it to gain more publicity.
Wonderful model that generates relatively good looking faces and prompts well. Of note, I do a lot of mixing with other models, using this both as a base, but also at the end as as refiner, and the model seems to play well with other models without much effort. It also plays well with LoRAs where necessary. Overall, great work, and I think it's a big improvement from the earlier versions I first used. Recommended!
Can you tell me any lora that enhances the skin texture more for this Checkpoint? That brings more realism? Thanks bro.
@null welp, it appears this person is now gone? Regardless, the answer is no, other than flat out mixing models. I find most photoreal models to be boring with their poses and angles, so it helps to inject a little anime at the frontend and then slap the photorealism back in.
Being a big fan from the early days, but all of the illustrious based versions so far suffer from a very noticeable lack of diversity and nationalities, where korean, african, persian, etc will almost always default to the same generic white girl (and the prompting isn't the issue).
The diversity of past models was some of its biggest strengths (especially XL and SD 1.5 ones). 3.5 doesn't seem to have made a dent. While I love what's been done with these in terms of realism, I really hope you work to add more variety in future versions. Illustrious and Pony have a huge issue with a heavy bias towards one look as anime / illustration first models, and they need lots of work to get them to a better state. Will keep an eye for future versions.
This model is based on Illustrious. As such, it also shares the 'limitations' of the base model. It's not intended to be an improvement on CyberRealistic 1.5 or SDXL. If you want the features you're referring to, you should use those versions instead.
If I modify CyberIllustrious to include those features, you’ll end up missing others that are specific to Illustrious.
it’s like making a recipe for a vegan cake, and then asking to include eggs and butter. You can do that, but then it’s no longer a vegan cake - and you lose the unique qualities of the original recipe :)
Very good model.
@Cyberdelia Do you recommend using ADetailer with this?
It depends - maybe not by default, but it can be incredibly effective.
Should we use your embeddings with this?
What embeddings? Character embeddings?
My absolute favourite Illustrious Checkpoint... But... for the love of god I cannot get it to produce Doggystyle POV Anal... Am I the only one? Anyone got tips?
It seems like you might need give it a bit of 'obedience training'! Try giving it with some clever prompting strategies, or perhaps it just needs more 'treats' in the form of fine-tuned parameters. With a little persistence and the right commands, you might just get it to 'roll over' and deliver that perfect 'doggy style' image you're aiming for! Keep in mind, sometimes you need to approach these things from a unique 'angle' to achieve the best outcomes ;)
Just merge it with a realistic sdxl checkpoint . Cyberillustrious, besides being the best nsfw illustrious model, is really great base for merging too.
@ElectricDreams thanks, merging my own checkpoints is a new concept for me. Do you have a good tutorial at hand you can point me to?
@TheAlfred756 if you are on comfy there are several checkpoint merge nodes from easy to complex. i don't know about a1111 but i think there is a plugin you can add if it's not in one of the tabs. it's really not that hard but it is time consuming because of loading times in between your testing. probably best is to use comfy so if you don't know how to use that yet, it is going to take a while to get used to but it gets more simple as you go.
I was always a bit skeptical about the Illustrious model being realistic, given its nature. But I have to say, this one is just awesome. I'm still experimenting with it, but the results so far have been great. That said, I'm having trouble getting it to produce low-key or dim lighting outputs. Also, it seems like there are no LoRAs available for the Illustrious realistic models to fix this. Any suggestions or workarounds for that?
This is a bizarre coincidence. I am working on creating a sort of nighttime LoRA for different versions to be able to make a sort of dark, night photo. I will also include Illustrious and see how that turns out.
And you are right. I have the feeling that CyberIllustrious is a bit of an underdog among the (CyberRealistic) checkpoints. It is not complicated to create good images. The just enough amount of NSFW and issues I had in the beginning are now resolved
@Cyberdelia that sounds awesome! I'm very looking forward to it!
@lordlexsus https://civitai.com/models/1508902
Maybe not exactly what you want?
@Cyberdelia I'll give it a try and let you know my thoughts. The example pics you provided look awesome already!
@lordlexsus SDXL, SD1.5 and Pony (maybe) will also be released soon. Same training data (266 professional photograph photos)
@lordlexsus You just didn't pay attention to this LoRA: https://civitai.com/models/623662/low-light-for-pony-variant - works great.
Here is an example of the author's render with this LoRA with the value =2
https://mega.nz/file/dRcClSyL#Cjahrnq3_W5ZamVaw8mA0DAVQsLNyjwTNHGwW_A40YM
It's also not bad, but it works differently:
https://civitai.com/models/915440/low-light-dark-chiaroscuro-for-illustrious-01-and-noobai
@windlike_gust It works okay-ish overall, but not quite in the way it should. Instead of adding realistic dim lighting (i.e., low-key ambient lighting), it simply overlays a somewhat dark, semi-transparent tint. Because of this, I wouldn't consider it a true solution for achieving authentic dim lighting effects.
Will it run in Forge UI?
100% :)
I use only Forge (reForge to be exact). Did you experience a problem?
Fantastic, great work, only downside is it hands out cleft chins like crazy.
Come on, the cleft chins aren’t Flux-tier tragic. 😄 Sure, a few sample images rock a cleft, but most have perfectly reasonable jaws… right? It’s weird though - same prompt, and sometimes you get a cleft chin, sometimes not. No logic to it!
I hope is not true, i don't use flux in order to not look at cleft chins (plastic skin doesn't help either). It's pretty much the last thing i wanna see in a sdxl checkpoint. I'm downloading rn, i'll see later on what is all the fuzz about
@ElectricDreams It's not like Flux that every chin is a cleft chin.
@Cyberdelia I've been getting it at about 80% of my images. But it's alright, between this checkpoint and others I have I can paint it out. Overall it's still an upvote.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.