
Join The Tinkerer on Whop to get early releases, private pages, and the Tinkerer Discord role - all in one place. š Join on Whop
CyberRealistic Pony Catalyst V4.0 (DMD2)
This is the special DMD2 edition of CyberRealistic Pony Catalyst, fully rebuilt for faster, ultra-realistic generations while using fewer steps.
Itās a major upgrade over the V3.0 release, delivering next-level quality for both SFW and NSFW creations.
Recommended Settings
Sampler: LCM
Schedule Type: Exponential
Sampling Steps: 8 (12 for extra fine detail)
CFG Scale: 1.5
ā” These settings are optimized for DMD2 and are not the same as my standard Pony workflow!
Patreon Exclusive Content
The full non-DMD2 version of CyberRealistic Pony Catalyst V4.0 is available exclusively on Patreon, along with other unique CyberRealistic checkpoints that are not published on Civitai.
In addition, future Catalyst models and experimental builds will also be released only on Patreon.
š Join my Patreon to access them all
Hires.fix Settings
Hires Steps: Can be the same or lower than your main sampling steps
Denoising Strength: 0.2 ā 0.5 (sweet spot for natural detail)
Hires CFG Scale: Keep it the same as your main CFG scale, or set to 1
CyberRealistic Pony Catalyst is the wild child of the CyberRealistic Pony family. Built on the foundations of CyberRealistic Pony but wired for experimentation, Catalyst isnāt here to play it safe. It's where bold ideas, unstable brilliance, and next-gen prompting chaos all collide ā and honestly? Thatās the fun of it.
This isnāt your usual polished release. Itās a testbed, a workshop, a what-if-we-just- moment captured in model form. Expect unexpected results, strange breakthroughs, and sometimes magic that might just become the next official evolution.
š§Ŗ What Makes It Different?
Built for exploration ā prompt freely, break boundaries, test weird stuff
Updated between mainline releases with no rules but vibes
Output can vary ā but when it hits, it hits hard
āļø Recommended Settings
Sampling method: DPM++ SDE Karras / DPM++ 2M Karras / Euler a
Sampling steps: 30+ Steps
Resolution: 896x1152 / 832x1216
CFG: 5
Clip Skip: 2š§ Prompting
Positive Prompt:
score_9, score_8_up, score_7_up, (SUBJECT), Negative Prompt examples:
score_6, score_5, score_4, (worst quality:1.2), (low quality:1.2), (normal quality:1.2), lowres, bad anatomy, bad hands, signature, watermarks, ugly, imperfect eyes, skewed eyes, unnatural face, unnatural body, error, extra limb, missing limbsscore_6, score_5, score_4, simplified, abstract, unrealistic, impressionistic, low resolution, lowres, bad anatomy, bad hands, missing fingers, worst quality, low quality, normal quality, cartoon, anime, drawing, sketch, illustration, artificial, poor qualityš§ ADetailer (Optional)
Adetailer model: face_yolov9c.pt
If you only want the main face being refined set 'Mask only the top k largest' to 1.š Prompt Support
Want better outputs? Use this custom ChatGPT to auto-generate elite prompts or fine-tune your own:
š [Click here for prompts]
ā Support the Mutation
If you like living on the edge of model development, fuel the chaos with a coffee. It helps me keep making weird stuff that works.
ā ļø Keep It Responsible
Catalyst can produce sensitive content. Be mindful of what you generate. Donāt be that guy.
Description
CyberRealistic Catalyst V2 is the realism-focused beta for the next generation of CyberRealistic Pony. Building on the flexible prompting style of Pony V10, this version brings back a stronger emphasis on realism - enhancing skin texture, lighting accuracy, and natural details.
FAQ
Comments (38)
Could you change the file naming convention to differentiate from CyberRealistic Pony? Something like "cyberrealisticCatalyst_vXY" would work.
Either way, great work, keep it up! ā¤ļø
Civitai renames it automatically.
Solution found (renamed version from V2 to Catalyst V2)
loving the extra realism of 2.0 but getting higher frequency of artifacts
Yet again looks amazing so with my tests so far
The new one is what v10 must have been. Put this into the main line.
And its awesome.
Itās rarely appreciated when I release a new model every week :) - which is exactly why CyberRealistic Pony Catalyst exists.
@CyberdeliaĀ How so?
@jamesluke609Ā Check out the comments :) For some reason, people get upset when I release new checkpoints too soon after the last one.
@CyberdeliaĀ Yep, who generates online, and old version dissapears:)
But you buzz this and the main ones too, no? How many models can you bid?
@jamesluke609Ā I pay over 50,000 buzz every week just to get my models into the Civitai generator.
@CyberdeliaĀ But, its worth it, isnt it?
@jamesluke609Ā Financially or just nice that it can be used? The latter is definitely true - the former is a bit disappointing :(
@CyberdeliaĀ Hmmmmm. I tought, you got the buzz mainly from online gen.
You are in top 5, if we count every model. So for less known creators, it absolutely not worth.
Yeah, Civit is sick and almost dying with self restrictions, so more users should learn local gen.
I hope, you stay.
@jamesluke609Ā Sure, Iām not planning to move to another platform. And for me, income isnāt the main priority (otherwise I wouldāve released this version under CyberRealistic Pony to make it instantly popular). Itās just a hobby that got way out of hand :)
i'm wondering why the pony scores couldn't be embedded into pony checkpoints.
This is how the checkpoint works. You can tweak the quality a bit using the weights. Embeddings, for example, can bundle those weights. I get the feeling itās not really a problem for you to keep adding them - you were just curious, right?
You can also just remove them. But some people will say you're not really getting the most out of the model that way. So, itās all about experimenting.
@CyberdeliaĀ hey Cyber, I'm a big SDXL user but I haven't used much Pony at all. What sort of Pony score lead-in should I be using? the entire string from score_9 to score_4_up or should I be stopping at 7 like I see you've posted in a couple places?
@shapeshifter83Ā Since this model is far from Pony base, you are good with usual positive(till 7), but try it with the whole strings(or without), you wont get much difference.
@shapeshifter83Ā added the usage on the info page of this model.
@CyberdeliaĀ Now I'm curious - why would you want a lower quality image? Is it because of variety in images?
@DrakPaiĀ It's a confusing science at times. Each token input will activate certain neural pathways inside the model relative to that token, but this always comes at the expense of the attention head strength of other tokens in the prompt. So it's not that anyone would want a lower quality image, it's whether or not a token like "score_6_up" is really worth the attention heads versus other tokens in the prompt. I don't know exactly what Cyber's training approach is here, but in general the more refined a Pony finetune becomes, the less "Pony-specific" the prompt will need to be since bit-by-bit the internal neural pathways of the model are less-and-less routed through the "score_#" gates.
Or it could be that Cyber is training with actual "score_#" tokens in the training so it could be more focused in that case.
Thus why Cyber is the one to ask, because only the one handling the training material and captioning would really know for certain, beyond someone simply doing trial and error to see what the outputs look like.
@shapeshifter83Ā Its more like XL and Pony hybrid with strong pony adherence. I can easily use XL loras(yes, the deleted celebs too) with some plus prompting, but for example Illustrious loras has little impact to this model(characters).
@DrakPaiĀ The simple answer is that a model needs to know what bad quality looks like in order to avoid it.
You can create or use already existing embed that does this for you. Altho many new checkpoints work fine without any scores.
I'm having such a blast with your GPT helper!
It's refreshing with a GPT that understands erotic and NSFW imagery, allowing reference images to be uploaded etc. Thank you very much for a fantastic service on top of your already great models! :)
I'm still fine-tuning it - glad to hear you like it! Iāve also created a GPT that helps generate wildcards.
@CyberdeliaĀ Hi there. I tried finding out what this GPT helper is or how to access it. Would you mind sharing or pointing me in the right direction? :)
@freakyfilemaster521Ā youāll need a free ChatGPT account to access and use any custom GPTs, even if you have the direct link.
@CyberdeliaĀ Thanks for the quick reply. I have a paid account so that will not be an issue. How can I try it?
@freakyfilemaster521Ā Ah ok :) The link is in the description. But here is it again:
https://chatgpt.com/g/g-6834133e3ab881918a91b3ec6b9eb01f-cyberrealistic-prompt-helper
Enjoy :)
@CyberdeliaĀ I didn't even realize it was a new feature - I just assumed I had missed it... š
Looking forward to testing it out with proper renders soon!
@CyberdeliaĀ I am not not sure how I missed that. It works great and thanks for helping :)
@freakyfilemaster521 @DrakPai I have added some sarcasm to the GPT if you ask a NSFW prompt. It's funny :) It still give the prompts like before but he is so tired of these NSFW prompts users are asking :)
Love working with it, it sure does have some "sass" to it.
Great realistic images!
Details
Files
cyberrealisticPony_catalystV20.safetensors
Mirrors
CyberRealisticPonyCatalyst_V2.0.safetensors
cyberrealisticPony_catalystV20.safetensors
cyberrealisticPony_v20.safetensors
cyberrealisticPony_catalystV20.safetensors
cyberrealisticPony_catalystV20.safetensors
CyberRealisticPonyCatalyst_V2.0.safetensors
cyberrealisticPony_catalystV20.safetensors
cyberrealisticPony_catalystV20.safetensors




















