CivArchive
    ✨ JANKU Trained + Chenkin & NoobAI + RouWei Illustrious XL ✨ - v4.0 🥇🔥
    NSFW
    Preview 75680746
    Preview 75680761
    Preview 75485065


    Tired of typing prompts + want viral-ready art?
    Pair Lazy Embeddings 🛋️⚡ (zero-typing magic for all models ) with TrendCraft 🪄🚀 (30-day trend-trained detailer ) for effortless masterpieces. Check them out: Lazy Embeddings + Trendcraft (Pony XL , Illustrious , Noob Vpred, or SDXL base trained version)

    👉 Join me on our discord server (https://discord.gg/tWrKk3jcXx)! Share art, ask for help, or just chat. We have CivChan bot plus games there!

    👉 Support future updates: Buy me a coffee or tip buzz! Every bit helps. Thank you for fueling the art! 🚀


    🚨 JANKU 🌟

    ✨ Key Features:

    • 🔥 LoRA-Free Workflow: Generate high-quality art without stacking detailers (unless you’re adding niche styles/concepts).

    • 🔞 𝐍𝐒𝐅𝐖Mastery:

      • Better anatomy (yes, even in complicated poses) + improved male framing.

      • 𝐍𝐒𝐅𝐖: if you want more control, put lazy𝐍𝐒𝐅𝐖 in negative!

    • Large character list support: Character list shared in our discord server.

    • No more stability issues —generate confidently at 1536x1536 without up-scaling. My personal resolution to use is 1024 x 1536.

    • 🎨 Crisp Details: Enhanced backgrounds, outfits, and character features.

    • 👀 Prompt Responsiveness: Gets your vision right the first time!

    • 🔧 VAE Baked In: No extra steps—just generate.

    • Read version details for the version you are using for more information!

    ✨Improvements from RouWei

    • Easy and versatile prompting with great aesthetic, anatomy, and stability

    • Vibrant colors, smooth gradients, and full brightness range

    • Over 35k artist styles pure training (no merged weights), and customizable options. The list in .txt can be found on our discord server.

      • Note: you must use by artistname to trigger it!

    • Eliminates watermarks, tag/character bleed, and strange artifacts for better coherence

    • Improved prompt following, stable styles, and accurate artist representation


    • Steps: 25-30 (I use 30 normally)

    • CFG: 3-5 (I use 5 normally)

    • Sampler: Euler / Euler A (my go-to)

    • Scheduler: Normal or Simple (my go-to) or SGM UNIFORM
      Model Recommended Resolution:768X1344 832X1216 768X1280 704X1408 1024X1536 (my go-to, no need for up-scaling!)

      • Standard Illustrious & SDXL resolutions also supported

    Prompting: Use my Lazy Embeddings for positive and negative and even 𝐍𝐒𝐅𝐖embeddings to prevent or add 🙈. Use lazywet to remove wet and shiny skin. No more guessworks for prompting! If you don't want to use embeddings, use your normal NoobAI/Illustrious quality triggers. There is artist, danbooru, and rating tag support.

    Positive prompt
    lazypos {character tags} {detail tags related to character} {background tags}
    lazypos, 1girl, makima \(chainsaw man\), yandere stare, smile, dress shirt, grey jacket, red hair, sidelocks, black necktie, looking at viewer, jacket, long sleeves, upper body, suit, necktie, shirt, simple background, medium hair, collared shirt, ringed eyes, shaded face, braided ponytail, head tilt, braid, white background, formal, yellow eyes, holding own arm, wing collar, closed mouth, solo
    Negative prompt
    lazyneg

    Pro Tips

    • Using my Lazy Embeddings for onsite gen doesn't cost you extra buzz, if anything it helps you save buzz since you're not generating bad images!

    • Check my images’ metadata for exact prompts/settings. Import them into ComfyUI to reverse-engineer my workflow!


    🩺 Fixes & Feedback

    • Eyes still wonky? Use eye detailer, but most of the time you don't need unless the subject is further away.

    • Found a bug? Report it—I’ll fix it in next version. I only generate with this model, so I’ll keep improving it. I won't publish something I won't use myself.


    ❤️ Why I Made This

    Tired of juggling 10 LoRAs for 𝐍𝐒𝐅𝐖 art? Me too. I want a 𝐍𝐒𝐅𝐖 checkpoint with more knowledge of characters and other common booru tags. This checkpoint is my personal playground —designed for creators who want bold, detailed, no-BS generations.

    Future updates? Always. Let’s push boundaries together.


    Please share your feedback to improve the checkpoint together! ❤️


    JANKU Checkpoint License & Terms of Service

    1. Comply with the NoobAI-XL 1.0 License

    2. User Responsibility :

      • You are fully liable for all outputs. Use prompts to filter content.

    3. Restrictions :

      • No redistribution of modified versions without adhering to open-source terms.

      • No use for unlawful or harmful purposes.

    Failure to comply voids usage rights.

    Description

    v4.0 Changelog

    • v4.0 is a checkpoint trained on thousands of images from popular and recent gacha games, primarily miHoYo titles. A100 GPUs were rented for training, as local hardware would take years.

    • After training, the model was further trained to remove low-quality concepts. By default, outputs are higher quality than without quality tags, and even more detailed with them.

    • Finally, a simple merge of my JANKU-style LoRA (trained on previous JANKU images) was applied. Merging too much risked losing trained knowledge, so only a small merge was possible.

    • Compared to v3.0, images are way more detailed (see below on how to adjust). A shift towards anime style while keeping the amazing lighting shades from v3.0.

    If you want less details or issues with hands, lower CFG to 3 and/or try scheduler Simple instead of Normal. Onsite gen works better with CFG 3, no clue why. If you use my positive embeddings, get the new lazypos and lazynegfrom the page. lazypos v1.0 is not good for v4.0. lazyneg v3 is optimized for JANKU. v4.0 is so good, you don't really need positive or negative prompts to start with, try "1girl" prompt and nothing else yourself and see the magic! 🔮

    👉 Join me on our discord server (https://discord.gg/tWrKk3jcXx)! Share art, ask for help, or just chat. We have CivChan bot plus games there!

    FAQ

    Comments (55)

    janxd
    Author
    May 11, 2025· 11 reactions
    CivitAI

    v4.0 is out in early release! Many hours of training went into this! Please read the version details for more information! Want to share your creations and chat with our awesome community? Join our discord! We have CivChan bot that helps share your art and Discord games with the community! Thank you all for your support! v4.0 won't have been possible without you all! Can't wait to see what everybody makes!

    renega51May 11, 2025· 10 reactions
    CivitAI

    can't wait to try!

    janxd
    Author
    May 11, 2025· 1 reaction

    Thanks so much for your support!

    WiafuThickerMay 12, 2025

    how is it,

    demonkingoftyrannyMay 12, 2025· 11 reactions
    CivitAI

    WOW, what an update. The accuracy is incredible, the model simply does what you want XD. You get satisfactory results immediately or very quickly. The sharpness and coherence of the background is also remarkable, not only that there are many clear and sharp details, the background also fits together as it is not mixed. The anatomy improvements are also clearly noticeable, especially the rate on 5 finger hands is quite high. All anime styles -> anime screencap, 2d, semi 3d, 3d, look very good. So far I haven't even got a single time too many or too few limbs, nor any strange poses. (The only thing is 6 fingers from time to time). A true masterpiece this could be the best checkpoint for anime. Thank you for blessing us with this.

    janxd
    Author
    May 12, 2025

    Thanks for your support! Your comment will be super useful for others!

    7721781May 12, 2025· 6 reactions
    CivitAI

    More Vibrant and richer Colors, natural shading is more accurate and prominent, fingers are 10/10 barely any mislimbed Hands and fingers, even the second Hand placemant is much better and less awkward when on other models. Prompts are working really well, less Need of reprompting certain things. the Background got alot more smaller Details. The increase in smaller Details (eye Details, accessoires, how clothing react under certain prompts, etc) Looks alot better as in v3. Also way less artifacts if you stack alot of loras, which still can happen but seems alot more resilent before turning blurry and artificial. Overall an 10/10 update can just recommend everyone to try it out, either now or at it's release!

    BingusChungusMay 12, 2025· 5 reactions
    CivitAI

    For those who've tried v4.0, is it enough of and upgrade over v3.0 to be worth buying some buzz to use it?

    janxd
    Author
    May 13, 2025· 2 reactions

    @Kalavith Posted a xyz plot comparing the two, you can take a look and judge yourself https://civitai.com/posts/16849074.

    I will try to bid for a spot for this model, so people can test with onsite gen. There is free runs for models in early access (if the creator offers it), so you don't have to pay yellow buzz first.

    7721781May 13, 2025

    Absolutely the v4 version is so much better over the v3! If you liked v3 you really should give it a go!

    cookiecatMay 15, 2025

    Doubling down on the improvement going into 4.0!

    Separate_WaysMay 13, 2025· 9 reactions
    CivitAI

    As always a spectacular job on this checkpoint. Your image quality is the best possible, if you are reading this you have come to the right place, this is the best checkpoint you will ever use.

    janxd
    Author
    May 14, 2025

    Thank you for your kind words! 😍

    mkthecollector614May 14, 2025· 6 reactions
    CivitAI

    Really good model that I every used

    cookiecatMay 15, 2025· 6 reactions
    CivitAI

    I already loved this in v3.0, but I have never felt such a HUGE leap in quality going into v4.0. I've found myself going back to old gens and being overjoyed with how it comes out on the newest version. Way to have my favorite checkpoint dethrone itself to become my favorite again.

    janxd
    Author
    May 15, 2025

    Thank you for your support! It will be very hard to beat v4 and might be the final version 😅

    janxd
    Author
    May 16, 2025· 10 reactions
    CivitAI

    v4 has onsite gen support now! There is free trials to test the model before purchasing with yellow buzz! Enjoy!

    WiafuThickerMay 20, 2025

    bro the image seems to be a little more paler in comparision to v3 any suggestion

    captainfutoMay 19, 2025· 10 reactions
    CivitAI

    Feet soles anatomy on the V4 is insanely good, actually the best I’ve seen so far! Huge thanks for the upgrade. Totally recommended! ❤️

    janxd
    Author
    May 19, 2025· 1 reaction

    Trained on creator preference 😏 Thanks for your support! ❤️

    ijghtdoiij480May 22, 2025· 2 reactions
    CivitAI

    dddd

    Silliest_MyraMay 22, 2025· 8 reactions
    CivitAI

    Genuinely this whole model feels fantastic! While I have had a few moments of goof, that's just my fault though. This is incredible otherwise!

    janxd
    Author
    May 22, 2025

    Thank you for your support!

    LYNX233May 23, 2025· 7 reactions
    CivitAI

    这是我用过兼容性最好的模型,当我使用许多模型都出不来我想要的效果的时候,它不会让我失望

    janxd
    Author
    May 23, 2025

    谢谢您的支持!❤️

    toshakinMay 24, 2025· 5 reactions
    CivitAI

    Hello, it seems my favorite LORA creator used your checkpoint, here's my result.

    https://civitai.com/posts/17374596

    Let me follow you and thank you so much, you made this possible.

    janxd
    Author
    May 24, 2025

    Thanks for checking it out! ❤️

    fvwevewvMay 25, 2025· 2 reactions
    CivitAI

    我是不是可以在这个模型上使用il和nai的Lora,或者是同时使用

    janxd
    Author
    May 25, 2025

    是的,它同时支持 ill 和 nai lora。nai lora 会更强大。

    fvwevewvMay 25, 2025

    @janxd 这个是基于il0.1的还是1.0\2.0的merge

    janxd
    Author
    May 25, 2025

    @fvwevewv noob eps v1.1 based on ill 0.1

    fvwevewvMay 25, 2025

    @janxd 可能是我对标题的理解有问题,我对标题的理解是这款模型是nai➕rouwei➕il的merge,如果我的理解没错的话我想知道这款模型merge的il版本,如果我理解错了请指正

    janxd
    Author
    May 25, 2025

    @fvwevewv Trained + NoobAI EPS + RouWei. ill 已经是 noob 的一部分,noob 是 ill 0.1

    janxd
    Author
    May 25, 2025

    不需要想太多,不管是 "ill" 还是 "noob" 的 loras 都可以正常工作

    fvwevewvMay 26, 2025· 1 reaction

    @janxd OK,非常感谢,期待你的下一个检查点

    janxd
    Author
    May 26, 2025· 13 reactions
    CivitAI

    v4.0 is out! Big thanks to all early supporters!

    Can't wait for everybody to be blown away by the huge improvements! Please share what you make, like always I will pin and tip the ones I like!

    Read version details for all changes for v4.0. If you use my lazypos embeddings, don't forget to get the newest version v2 to use with JANKU v4.0.

    If you have a lot of blue buzz you don't use since you generate locally, please generate some onsite to show me some support!

    If you haven't already, please join our discord for updates, chat, share art, and play with CivChan bot!

    For onsite gen users, after this week I will only bid on v4.0 as it is way better, so it doesn't make sense to keep both version up. If this is a problem for you, let's talk in discord!

    It is also on tensor art https://tensor.art/models/867657363060974134

    janxd
    Author
    May 26, 2025

    Thanks for the support @minging 💪

    Fujin00May 26, 2025· 8 reactions
    CivitAI

    It's definitely a great update, but the contrast and color saturation are way below the previous one.

    janxd
    Author
    May 26, 2025

    did you try adjusting the sampler/scheduler and cfg?

    Fujin00May 27, 2025

    @janxd Although I always used the same settings, I was uncomfortable with this color palette and changed my settings, but the result is the same.

    janxd
    Author
    May 27, 2025

    @Fujin00 Will need to adjust between v3.0 and 4.0 in terms of settings, can't use the same. If you can share some pictures or come to discord we have people that can help you get the colors back!

    BingusChungusMay 27, 2025

    Have you been able to fix this? I am having the same issue, every image is overexposed.

    Fujin00May 28, 2025

    @BingusChungus No, I couldn't fix it. I'm trying to get used to using it like this.If you find a solution, I would be happy if you let me know. :)

    Fujin00May 28, 2025· 1 reaction

    @janxd It's not annoying, it's very little. I guess it's just noticeable because I'm used to the previous version.

    janxd
    Author
    May 28, 2025· 1 reaction

    you can try swapping the vae to https://civitai.com/models/1290283/neptunia-xlilnai-vae?modelVersionId=1455983 but I don't have the issues you two are having so using this makes the color have too much constrast/color saturation

    BingusChungusMay 27, 2025· 5 reactions
    CivitAI

    While I like v4, I feel like the contrast and colors are worse, almost every image feels a bit overexposed.

    I do think the prompt adherence and anatomy might be better, but the main thing I noticed is the above. I believe I will stick with v3 for now.

    janxd
    Author
    May 27, 2025

    share some images please along with prompt and settings! also people in our discord server can help too!

    BucchaMay 28, 2025· 12 reactions
    CivitAI

    ♥Your checkpoint v4.0 turned out really well. Thanks for that. ♥

    janxd
    Author
    May 28, 2025· 1 reaction

    @Buccha Thanks for checking it out! Love what you made with it so far!😍❤️

    mkbmbt126com649Jun 1, 2025· 6 reactions
    CivitAI

    Great work and thx for sharing. There's a small issue here and I'll be very grateful if anyone can read this. I use RTX3070 to generate pics now. When I use other checkpoint (like WAI's and Hassaku's), it is enough to generate 1536*1024 basic pic and upscale to 1920*1080 in the meantime. But When I switched to this one with same generation settings, it reports a out of memory error. Does this model requires more GPU resources or I need some additional settings for this model (I already did as the instruction)? THX!

    janxd
    Author
    Jun 1, 2025

    If you are switching models a lot, it is because the vram didn't clear. Kill the application or unload the model from vram if you know how and start it again and make sure vram is cleared.

    mkbmbt126com649Jun 1, 2025

    @janxd Thx, I didn't notice the memory usage of GPU. I think I should upgrade my device as soon as possible. It seems GPU with 8GB memory is not enough now.

    janxd
    Author
    Jun 1, 2025

    @mkbmbt126com649 8gb should be enough as we have users in discord using 8gb. Just need to make sure it is cleared from other apps as well!

    mkbmbt126com649Jun 1, 2025· 1 reaction

    @janxd Thx, it works fine. The outputs are amazing.

    janxd
    Author
    Jun 3, 2025· 10 reactions
    CivitAI

    If you use my lazy embeddings, lazyneg v3 just released: https://civitai.com/models/1302719?modelVersionId=1860747 Optimized for JANKU v4 🤘