If you like my work, drop a 5 review and hit the heart icon. it's free and keeps me motivated
All my models are officially hosted and maintained by me on Tensor.art . use my Exclusive and public model for free on tensor.art
v5.0 civitai onsite generation enabled : 04/01/25
v5.0 final release : 01/01/25
v5.0 pre relelase update : 25/12/24
Check tensor.art for V1,V2,V3 online generation support
You can run Prefect Pony XL and use its API on SinkIn
currently, only v5 and v4 are available for civitai generation.
v3,v2,v1 are retired from civitai generation. check other version details for alternative on site generation service.
WIP: Realistic
V5 Update :
onsite generation : https://tensor.art/models/815661199413897287
pre-release : https://tensor.art/models/810892802167877454
Added trained Lora
Cleaned style
Baked in VAE
Extra trained data of 15000 steps
Better Lora support.
V4 Update :
onsite generation : https://tensor.art/models/794513755405953653/Prefect-Pony-XL-v4
Added trained Lora
Cleaned style
Baked in VAE
Extra trained data of 15000 steps
Better Lora support.
V3 Update :
onsite generation https://tensor.art/models/772139385270444718
added some style to it
should give better face and eyes
better NSFW supports
wide range of Lora support
V1 lightning :
8-steps
DPM++ 2M
V2 update:
onsite generation https://tensor.art/models/754887012743423435
has less style influence
can work with wide range of loras
better anatomy
better color
V1 update:
On site generation closed for v1 on civitai for limitation of civitai.
try tensor.art if you want to use the online generation for v1
https://tensor.art/models/724940954567461228/Prefect-Pony-XL-v1.0
If you like my work then drop a 5 review and hit the heart icon. it's free and keeps me motivated
Suggested settings:
I had CLIP skip 2 on every image
I had ENSD: 31337 all of them
All of them had highres.fix or img2img at higher resolution.
I don't use restore faces
I use afterdetailer for face details
Tiled diffusion for img2img upscaling and sometimes Noise Inversion for some more extra details.
4x-Ultrasharp upscaler
NOTE: if you find any prompt of the preview images look familiar it's because I've taken them from other model review images, Credits to the original authors. Thanks for the benchmark.
Commission open in Patreon.
Get early access and Exclusive NSFW Lora in my Patreon .
Support my work by joining any one of them and get early access to all my upcoming loras and other perks such as fan requests and Discord role.
Join my Discord Server
Description
On site generation closed for v1 on civitai for limitation of civitai.
try tensor.art if you want to use the online generation for v1
https://tensor.art/models/724940954567461228/Prefect-Pony-XL-v1.0
FAQ
Comments (80)
not sure if the website is screwing up again or if there are no preview images...
all are mature image so that might be causing the issue.
I'm making more sfw and hoping they wont be flagged by the ai
when goofy makes a pony, people should pay attention
this guy was right ^
Perfect and Very Beautiful ! Thank you !
Delicious! Amazing model 💋
When I try to use this checkpoint, all I get is noise patterns - any idea why?
Try removing all negative first . Also check the vae. Either change it to sdxl or don't use any vae
https://civitai.com/articles/4044 Read this.
@ABSDE Thanks but unfortunately this didn't help as it claims it doesn't happen in Comfy, which is what I am using, and thee is no fix listed for Comfy.
@PodRED Sad, because I use the vanilla Automatic 1111 (which is way easier to use than comfy, if I were to reinstall webui one day because something real bad happened I migrate to Forge)
The model is insane but im having serious problems to make futa characters unlike with others models, hope you can fix it on 2.0
Not sure about you but I'm getting pretty good futa girls. Maybe check the prompts and see if other resources are conflicting or not
@Goofy_Ai Well I have no idea, I've tried the same fairly simple prompt (1 girl, futanari, large penis), in several models and in yours I almost never get a futanari, I've tried a thousand times and with all samplers and I do not know what can happen, but it is definitely a problem with the model and the futa girls, if you notice also, there are hardly any futanari photos in the images that people have uploaded with the model, which is quite rare because it is a very popular topic, it's a shame because the aesthetics of the model is amazing.
just a general question if anyone could help.. how do you control the height of the character? and what prompt can you use for change camera angle from high (like looking down) to low(looking up)
beatiful 💗
The best Pony model I have used. Thanks a lot!
Nice model, only a thing, from your settings:
"I had CLIP skip 2 on every image", stable diffusion xl hasn't clip skip so no difference with clip skip 1 or 2 with sd xl
well, its a pony xl based model and the pony model author mentioned using clip skip 2 so that is what I'm using and what I'm suggesting.
"Important information
Make sure you load this model with clip skip 2 (or -2 in some software), otherwise you will be getting low quality blobs "
@Goofy_Ai With webui 1111 no difference I tried with your model and under X/Y/Z plot CLIP SKIP 1,2 the two images are exactly the same... it's a small thing, nice model nothing to say
@shiroppo thanks but its better to be safe. the model is being used in different platforms and people do use different UI too. no harm is staying in the safe side by following the original settings.
@Goofy_Ai I did some checking... probably with older versions of webui 1111 there was a bug(different clip skips with sdxl caused different results when applying one or more lora)... better to be safe indeed
isn't this autismmix just renamed, also your license is incorrect, you're supposed to use whatever is set for them when using either pony or animagine
🦫
good
What is ENSD: 31337??
extra noise seed delta i think and people used to set it as a parameter back in the automatic1111 early days for i think more consistency when sharing prompt images to be recreated by users,
but could be wrong.
Best Pony checkpoint... thank you for sharing..
What are the upscaler settings supposed to be? I can't get it to upscale without deforming the picture like crazy
use denoising strength of less than 0.3
very nice implementation
Why am I only getting hyper realism pictures and not anime styled? Even copying other users prompts it only spits out realism
Hard to guess specifically what problem you're having when you say absolutely nothing about how you're trying to use the model...
Civitai's generator? Something web based elsewhere (which?)? Did you download the model and run it locally (using what?)? What kind of settings are you using?
If you aren't getting anything anime, specifically put 'anime' or 'anime style' or something similar in the prompt. Put it near the beginning or add emphasis to it. I can assure you that this model definitely can create anime-style images, and at least in my experience I didn't have to work for that at all, it more-or-less defaulted to it.
It is possible that you aren't using the same embeddings such as PDXLNegv2, etc. Are you using a refiner? This could also be causing issues. I have had some issues with hyperrealism photos in PDXL despite declaring source_anime or source_manga. I include hyperrealism in my negative prompt.
Can you do a LoRA of Jackerman style? If you do that would be great 👍 🙏🙏
Since my computer is too weak to do XL and Pony, I've been using the generator and I have to say, this is arguably my favorite model to gen with. Thank you and keep up the good work
try creating lora dragon
Very solid, reliably nice results!
oh no, not "prefect" xD
Amongst all the latest ponys and add-ons for it, This one stands out, being flexible on its own and can create beautiful results with just prompting, I usually img>img my pics from pony to another model for enchantment, but since I cant do that on this page, this one does the trick and more so!
Great stuff!
这个模型什么都好，就是不太“听话”~
very solid sheesh
This model is goated.
this model is f*cking amazing!
work furry?
Solid model, I use this one often as well a couple others.
Can i know what is the recommended base resolution before using hires with ultrasharp upscaler for 16:9 aspect ratio?
Very frustrating to use if you want to make femboy/otokonoko/trap stuff, because no matter how you phrase it or how many different ways you put "female" or "girl" in the negative prompt, it makes everything female regardless. I've even done (female:99) and it still gives me females. I'm guessing it wasn't trained on femboy images....
It works sometimes, and it's great when it does, but 95% of the time, it just makes females even when you tell it not to and it's absolutely infuriating.
idk about you but i just took a random gelbooru tag set from a "trap" image and it gave me a femboy on my first try on civitai generator
@Goofy_Ai Yeah, works just fine for me for traps/femboys. The key thing is (bulge:1.5), should almost always give good results.
@Goofy_Ai Yeah, like I said - it works sometimes. Also, I'm not using the civitai generator, I'm using automatic1111.
Amazing
Wow
I'm in love with this model, it can cook female figures very well. Beautiful art style I can say. Could you please let me know what is this art style called? Or which lora styles did you use for this model?
it works really good (translation: makes my peepee go zooweEEEE!!!)
coming back 4 months later to state that it still works really good (translation: Goofy_Ai deserves at least 5 big bitches not even kidding)
Best Pony model. i testing on 30+ image. this checkpoints have minimum errors, dublicates, and + fingers. I wiil be training on this new lora.
Great generations and very easy to use. I hope to see the same quality moving forward, especially when this BASE model gets updated.
I have everything over-lighted with this model. It's like the bright light isn't adjusted. Why? Auto1111
@Goofy_Ai
What model and settings do you use on adetailer? I can't get it to do anything.
you can check the image metadata for all the settings. i normally use the default settings
@Goofy_Ai default would be 0 then, the OP is referring to the adetailer de-noise scale amount and other settings on adetailer, but aside from using the ultrasharp upscale model, you used 0 on everything?
So far, it is decent. Not perfect, not too horrible.
hey! newbie here, is this version better/latest than https://civitai.com/models/257749/pony-diffusion-v6-xl?modelVersionId=290640
which one is better/latest?
the link you provided is the base model. and the prefect pony is a model that is a merge of that model
@Goofy_Ai I see! Thanks for the info
It looks like the original v1 file was uploaded as v2 (same file,same sha256: 91b87bf726bf35bc7bdbdac1656102b1c9131df0787c0611cc4c0ad93809d858)
my dumbass missclicked.
slow internet . uplaoding
@Goofy_Ai Happens to the best of us. But thanks to this, I now am aware of your new model. Can't wait to test it out, Prefect is one of my favourite models!
@LockMyHeart fixed
@Goofy_Ai thanks, I'm actually curious to see it's difference between v1 and v2, I'll post images soon of comparison.
@Goofy_Ai Sweet! And I see it's even up on the on-site generator, so that's great!
Details
Files
prefectPonyXL_v10.safetensors
Mirrors
prefectPonyXL_v10.safetensors
prefectPonyXL_v10.safetensors
prefectPonyXL_v10.safetensors
perfectPonyXL_v10.safetensors
prefectPonyXL_v10.safetensors
perfectPonyXL_v10.safetensors
prefectPonyXL_v10.safetensors
prefectPonyXL_v10.safetensors
prefectPonyXL_v10.safetensors
prefectPonyXL_v10.safetensors
perfectPonyXL_v10.safetensors
perfectPonyXL_v10.safetensors
prefectPonyXL_v10.safetensors
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.