CivArchive
    Preview 61941732
    Preview 61941730
    Preview 61941727
    Preview 61941719
    Preview 61941726
    Preview 61941706
    Preview 61941725

    Anima Version is Available!!

    Nova Anime XL

    Nova Anime XL is Nova Anime: Anime/2.5D/3D checkpoint model for SDXL

    Pony version aims to get the similar look as both Nova Anime and Nova Domain

    All images on example were created with diffusers with custom png tags

    SeaArtAI

    OurtAI

    PixAI - IL v19.0

    PixAI - IL v18.0

    PixAI - IL v17.0

    PixAI - IL v16.0

    PixAI - IL v15.0

    PixAI - IL v14.0

    PixAI - IL v13.0

    PixAI - IL v12.5

    PixAI - IL v11.0

    PixAI - IL v10.0

    PixAI - IL v9.0

    PixAI - IL v8.0

    PixAI - IL v7.0

    PixAI - IL v6.0

    PixAI - IL v5.5

    PixAI - IL v5.0

    PixAI - IL v4.0 Happy Valentine

    PixAI - IL v3.0 Happy New Year

    PixAI - Illustrious v1.0

    PixAI - Illustrious v2.0

    PixAI - IL v2.5 Merry Christmas

    PixAI - Pony v1-v2

    PixAI - Pony v3

    PixAI - Pony v4

    PixAI - Pony v5

    PixAI - Pony v6

    PixAI - v7.0

    Rules

    You cannot use the generated images for commercial use if it's not edited (or just turning it to black and white)
    You can share images without any restriction if you don't monetize it

    Advertising this model to outside is always welcome

    Recommend Settings

    Sampler: Euler a

    Steps: 20~30

    Clip Skip: 1-2

    Denoising Strength: 0.65 - 0.8
    (Pony)

    CFG Scale: 5~7

    Prompt: score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, source_anime, BREAK

    Negative Prompts: score_4, score_5, 3d, jpeg artifacts, username, watermark, signature, normal quality, worst quality, large head, low quality, text, error, missing fingers, extra digits, fewer digits, bad eye

    (Illustrious)

    CFG Scale: 4~6
    Prompt: masterpiece, best quality, amazing quality, 4k, very aesthetic, high resolution, ultra-detailed, absurdres, newest, scenery, {Prompt}, BREAK, depth of field, volumetric lighting

    Negative Prompts: modern, recent, old, oldest, cartoon, graphic, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured, long body, lowres, bad anatomy, bad hands, missing fingers, extra digits, fewer digits, cropped, very displeasing, (worst quality, bad quality:1.2), bad anatomy, sketch, jpeg artifacts, signature, watermark, username, signature, simple background, conjoined,bad ai-generated

    Description

    Original: IL v5.5

    Have the looks of v3.0 and more contrast
    NoobAI EPS 1.1 applied

    FAQ

    Comments (19)

    hushichoMar 7, 2025· 11 reactions
    CivitAI

    5.5 definitely isn't more stable than other models, maybe because of what was added. Unusable.

    Crody
    Author
    Mar 7, 2025

    Although the name is v5.5, it's alternative version for v5.0 since that one feels nothing like "Anime"
    So, this model exists as Anime coloring + Flat posing model
    The most stable model I made would be Nova Orange but that one is too stable to make it as Anime

    Noxy_Mar 7, 2025

    @Crody That does make sense and funnily enough, I agree with the above user, it's definitely unstable, as it never listens to my negative prompts now and even worse anatomy

    EverghastMar 7, 2025· 11 reactions
    CivitAI

    Hello, as I really love your models, I'm just giving you my feedback for the 5.5. The 5.0 was working perfectly, colors were excellent, prompt but the newer version has color saturated a bit too much, less detailed it seems, in my generations compared to the 5.0 and a bit less stable, as someone said before in the previous comments. About the prompt adherance, I am not sure either: characters elements such as clothing or skin colors are not as well distributed than in the 5.0 At least, this is about my experience, perhaps it is me doing shit, but as it works in the previous version, i prefered to comment here. :)

    PalmtopTigerMar 9, 2025· 14 reactions
    CivitAI

    Hi. Nova Anime v5.5 breaks some sampling methods. For example, Euler (non ancestral) and DPM++ 2M give noisy results with many artifacts. The same problem can be observed when using Nova Furry v4.0.

    Crody
    Author
    Mar 9, 2025

    The model is fine tuned and added some detail loras to get the result
    Because of that the image gets noisy when you use samplers with great details
    I suggest using Euler A with 20 steps, CFG scale 4.5

    raffyffyMar 9, 2025· 10 reactions
    CivitAI

    im loving the lower contrast on 5.0, it makes it look more anime in my opinion <3 love it

    halukaleeMar 9, 2025· 12 reactions
    CivitAI
    v5.5 cannot work normally, the author should specify the parameters. Using the parameters of V5 and versions below V4 will result in serious noise. According to the parameters recommended by the author in other user comment sections, CFG4.5, Euler, 20 steps, the noise is still severe and cannot be alleviated. At least, it does not work normally in comfyui at the moment.
    Crody
    Author
    Mar 9, 2025

    Probably because I use hires.fix at x1.5, 0.4 denoise and 40 steps
    If 4.5 CFG gets noisy, try using between 3 - 4 instead

    gausssidorov928Mar 9, 2025

    @Crody  Тут стали получатся уродливыми части

    UrashimaTaroMar 14, 2025

    I had the same problem, you need to set the sampler to Euler Ancestral, cannot just leave it on default.

    halukaleeMar 14, 2025· 1 reaction

    @UrashimaTaro yes,it worked.thank u.

    Noxy_Mar 9, 2025· 16 reactions
    CivitAI

    As other commenters have said, V5.5 is unstable and for me it never listens to my negative prompt sadly

    That said, V5 works for me and it's decent

    Edit: I'll be switching to V3 since it seems like V5 has the tendency to either listen or not, too inconsistent of prompt adherence after a bit of a few days of more use (V5A or V3 is my recommendation if you don't like how the faces look in V4, like me)

    civitaithrowaway28398Mar 11, 2025· 16 reactions
    CivitAI

    These are the best models I've used since the OG Pony stuff. Yowzer!

    316908Mar 12, 2025· 14 reactions
    CivitAI

    With new on-site generator update, I finally can use my favorite checkpoint to gen from everywhere now! Thank you for your great work!

    Wa1nukeMar 13, 2025· 16 reactions
    CivitAI

    Best model, actually❤️

    SargoxMar 18, 2025· 20 reactions
    CivitAI

    IL v4 is currently the most stable version. IL v5A has very low contrast, which needs to be enhanced either through external LORAs or by increasing the CFG to 7. IL v5B exhibits excessive noise and over-detailing, even with Euler A. IL v5B produces poor results with the same parameters that worked well on IL v4.

    aigenerator456Mar 26, 2025· 18 reactions
    CivitAI
    thantuongso50Mar 31, 2025· 16 reactions
    CivitAI

    Great Model, NGL <3