Sign In

Reviews

Animagine XL 4.0

Overwhelmingly Positive

99% · 3,529 reviews

Zap420's Avatar

Zap420

Purple Panther
· v4 Opt
Recommends

best

0 comments
Read review by Zap420
cicil1990's Avatar

cicil1990

· v4 Opt
Recommends

really good

0 comments
Read review by cicil1990
Recommends
THE ABSOLUTE BEST when it comes to dynamic poses, no more stiff poses or characters posing for the camera, more loose, more fluid the only issue is hands, sure u get the wrong anatomy here and there but that's a given when looking for unusually poses but the hands are always wrong, i wish for an update to fix only the hands
0 comments
Read review by yokinarudesu351
VexenAI's Avatar

VexenAI

· v4 Opt
Recommends

This is one of my favourite checkpoints. Great for multiple character LoRAs in one image

0 comments
Read review by VexenAI
Doesn’t recommend
sry but this version kills my ai and anime heart i tryed few gens and i cant rly get good results sure it understand prompts but mostly wrong wtf does slender beutiful eyes etc with the char bruh it kills me an model improvement with depth of field? kills me too if u want a all over good model pls dont use depth of field and check the prompt understanding , cuz specialy me uses beautiful face or eyes + best quality .... ty
0 comments
Read review by schneesturmx91988
Fontose's Avatar

Fontose

· v4 Opt
Recommends
A good semi-realistic anime model. Fit well colors, ambient lightning and background, making result which want to stare at details for a long time.
0 comments
Read review by Fontose
bolteg's Avatar

bolteg

· v4 Opt
Recommends

Good **** brotha

0 comments
Read review by bolteg
BA

BACCIV01

· v4 Opt
Recommends
It does execute what I prompt without problems, compared to the 3.0 and the 3.1 versions.
0 comments
Read review by BACCIV01
Lilium_Album's Avatar

Lilium_Album

· v4 Opt
Recommends

When compared to 3.1 it gets significantly better backgrounds and characters can be generated with incredible accuracy and minimal description. However I don't like that it is seemingly extremely dependent of following the recommended CFG and sampling method, any variation on these parameters will result in significantly worse results. The updated Dataset is much welcome, but I am not much of a fan of the seemingly strong bias towards generating pictures with a cell shading technique, this can be worked around with enough prompt expertise but one of the reasons I loved 3.1 was what they call the "2.5D aesthetic" which was actually a side effect of over-using the aesthetic-score keywords, but the fact is I still love it, and it is much more difficult to get with this version. That being said, this one definitively feels like an improvement and a technological advancement. I would like to Thank the kind people of Cagliostro Lab for all their hard work and for sharing this lovely checkpoint to the public for no cost at all

0 comments
Read review by Lilium_Album
ShinyLilly's Avatar

ShinyLilly

· v4 Opt
Recommends

impressive model for characters and art!

0 comments
Read review by ShinyLilly
Recommends
Good.
0 comments
Read review by c4rt4philus179
Recommends
Fantastic model, support over 200 styles, tons of characters without using any lora. Easy to use prompts system etc.
0 comments
Read review by thamngapwei546
MG

MGK

· v4 Opt
Recommends
a nice model to paint anime
0 comments
Read review by MGK
GR

GreatMonster

· v4 Opt
Doesn’t recommend
Although Animagine XL 4.0 is good at replicating artists' styles and can relatively well reproduce those with limited datasets, it has significant issues with image quality, composition and anatomy. Elements tend to blend together similar to SD1.5, and the recommended negative prompts seems to ruin the image in my test.
0 comments
Read review by GreatMonster
MintXil's Avatar

MintXil

Elemental Extravaganza Contest Participant
· v4 Opt
Recommends
confirming much more consistent outputs, very stable generations without using any Lora and/or Embeddings - just love to experiment! thanks for this ♥
0 comments
Read review by MintXil
S27's Avatar

S27

Project Odyssey Contest Participant
· v4 Opt
Recommends
Fastest see model>click>download in all of my civit browsing history. Haven't played with it too much yet but already with the usual negatives and basic prompts getting some excellent results! Seems like a worthy successor, GJ!
0 comments
Read review by S27
plsdonotdisturb's Avatar

plsdonotdisturb

· v4 Opt
Recommends
single-handedly the best anime model since using it from v3 :D
0 comments
Read review by plsdonotdisturb
MA

maxnami517

· v4 Opt
Recommends

"No need" for Loras.

Using the correct tags that they used from training would deliver a nice generation without any loras help.

0 comments
Read review by maxnami517