CivArchive
    Preview 84602807
    Preview 84602795
    Preview 84602802
    Preview 84602810
    Preview 84602791
    Preview 84602800
    Preview 84602804
    Preview 84602806
    Preview 84602794
    Preview 84602801
    Preview 84602790
    Preview 84602797
    Preview 84602809
    Preview 84602811
    Preview 84602805
    Preview 84602803
    Preview 84602808
    Preview 84602788
    Preview 84602793
    Preview 84602798

    I believe this model to be very good at photorealism although it does have some artistic capabilities as well. Here are a few tricks to using the SDXL versions of GonzaLomo to its fullest potential:

    • Clip Skip is your friend. I recommend trying a prompt with both 1 and 2 values. 1 can make images that are more creative, artistic, retro or grungy whereas 2 is better at photorealism and the, um, physical attractiveness of the subject(s).

    • Generate at 1024x1536 or higher using the PatchModelAddDownscale node in ComfyUI. This will create a much more detailed image that can then be upscaled further, resulting in incredible photorealistic detail. Using that node, I can do an initial generation as high as 1280x1920.

    • You can play around with different sampler settings but I find that the absolute best results are with LCM/Karras, 12 steps and CFG 1.0.

    • Selective usage of the Self Attention Guidance node can sometimes give amazing results. I always have it in the pipeline and usually bypassed but turn it on when I think an image has potential but just needs a little more oomph like color saturation or contrast.

    • I get really interesting/creative results by generating with the V1.1 Semi-Real model and then resampling that image with the V2.0 Unity model at around 0.60 denoise. You really need to play with the denoise to find the right value that tips the image from cartoonish to photorealism but you'll be amazed at what's possible when you do. Note that V1.1 Semi-Real requires Clip Skip 2+ and will not work with 1 at all.

    • And lastly, always be upscaling! More pixels equals more detail equals more photorealism.

    Note: I only use ComfyUI so I apologize in advance if you're having trouble with my model in some other tool and I can't help.

    Current Models

    V2.0 Unity

    Essentially, this is V1.1 Real with a bit of Siren Forge, Cyberrealistic and Lustify added to the mix. The base is still Gonzales and LomoXL.

    I called this version Unity because I originally wanted to have separate models for photorealism and artistic styles. However, I realized that I could get both in one model and created V1.1 Artsy for that purpose but it was a failure. This is my second attempt at that combination and I think/hope it's a success.

    At the very least, I don't want to degrade the photorealism capabilities and just want to add more surreal, artistic or fantastical styles as an option.

    The recommended settings haven't changed since V1.1 Real:

    Sampler: LCM (also dpmpp2a and ddim can provide interesting results)

    Scheduler: Karras, Exponential or Beta

    Steps: 8-12

    CFG: 1.0-1.3

    Clip Skip: 1-5

    I normally keep my settings at LCM/Karras with a CFG of 1.0 but I tend to vary the clip skip between 1 and 5 because those can change the image pretty drastically. Sometimes an image at clip skip 2 will be terrible but might become amazing with clip skip 1, and vice-versa.

    V1.1 Semi-Real

    I'm doing a little experimenting with non-photorealistic models. This is a merge of GonzaLomo v1.1 Real and ToonComix which is a toon model that does a really good job with detailed backgrounds. The whole Comix family of models is very good so check them out.

    Sampler: LCM

    Scheduler: Karras, Exponential or Beta

    Steps: 8-12

    CFG: 1.0-1.3

    Clip Skip: 2-5 (1 is not supported)

    Retired Models

    V1.1 Real

    This is a simple 60/40 merge between two of my favorite photorealistic models - Gonzales-NSFW-PonyV1/V2-DMD v2.0 and LomoXL (with a little bit of Lustify thrown in). Many, many thanks to the makers of those excellent models.

    Gonzales is a Pony model so GonzaLomo inherits all of those capabilties. LomoXL has a very strong analog feel to it which helps give an edge to Gonzales' "pony-ness".

    Sampler: LCM

    Scheduler: Karras, Exponential or Beta

    Steps: 8-12

    CFG: 1.0-1.3

    Clip Skip: 1 allows greater analog effect but lessens NSFW capabilities

    Clip Skip: 2-5 lessens/removes the analog effect but allows more NSFW

    Try this workflow to get started, if you need it.

    To be clear, I created this merge for my own use cases. It's certainly not better than Gonzales for most things. I'm just hoping to inject Gonzales with some of the gritty, analog feel that I was looking for in some cases. LomoXL also allows for a wider variation of faces, especially with clip skip 1.

    Enjoy!

    P.S. This is my first model upload so I would appreciate any constructive feedback or comments.

    V1.1 Artsy alpha (Not very good imo, use V2.0 Unity instead)

    This was an attempted merge between V1.1 Real and a few art-focused models. The main purpose of this model is to be able to produce artistic styles but it has the tendency to try to force photorealism when prompted clearly for an artistic style. This model has been abandoned in favor of V2.0 Unity which does a much better job of combining the photorealism of V1.1 Real with artistic styles.

    Description

    Added a bit of Analog Madness and Epic Realism to the mix and really liked the result after some tweaking so hopefully you like it too. V3.0 seemed to be very popular so my main goal is to not mess anything up with its capabilities and just provide an alternate version.

    The recommended settings are the same as previous versions - LCM, CFG 1.0 and 8-12 steps. However, instead of Karras as the scheduler, I actually think this version works better with Exponential. Karras still works great too though, as does Beta.

    FAQ

    Comments (20)

    realismenhancer585Jun 26, 2025· 4 reactions
    CivitAI

    Excellent! The new version looks very promising — I’m excited to start testing it right away. Thanks for your hard work, I’m really looking forward to seeing the results! ❤️

    GBRX
    Author
    Jun 26, 2025· 1 reaction

    Thanks, I hope you like it!

    7581535Jun 26, 2025· 2 reactions
    CivitAI

    Great! My first tests seems to show better prompt adherence, Especially for short prompts. Thanks a lot!

    GBRX
    Author
    Jun 26, 2025· 1 reaction

    Thanks so much for verifying that. I had thought my prompts seemed to work better as well but I didn't want to promise improved prompt adherence. I do think longer, more natural language prompts don't work as well with any of my models unfortunately.

    7581535Jun 26, 2025· 1 reaction

    @GBRX in many instances the adherence gets messed up by the choice of portrait or landscape as the model can't fit things horizontally or vertically. That's mostly true for full prompts wildcards.

    AstelteynJun 26, 2025· 6 reactions
    CivitAI

    You know, usually if they make a good model, I rarely expect improvements in new versions, as a rule, the opposite happens, the new version is worse than the old one. But here everything is good, version 4 is really better overall than version 3, I tested it for about 7 hours with different setups, overall it is really better if you take the average statistics.

    GBRX
    Author
    Jun 26, 2025· 2 reactions

    Awesome! I really appreciate you doing all that testing. Of course I did some myself and pretty much came to the conclusion that I thought the overall results were better even if they were worse in a few instances. That said, I kind of think of v4.0 being an alternate version to v3.0 rather than a huge improvement but I do prefer the v4.0 images on average.

    BinaryCanvasJun 26, 2025· 4 reactions
    CivitAI

    Very nice checkpoint... again! Thanks!

    ElectricDreamsJun 26, 2025· 7 reactions
    CivitAI

    I've noticed that latest sdxl checkpoints are really pushing the limits of the model. I mean, while feats like prompt adherence are still quite basic, the IQ is really high at this point. Not only in this checkpoint but a few more. But in a few days, those few checkpoints will be many. I personally think we're way beyond photorealism, i don't know if you share this insight. Do you reckon there's much to improve regarding image quality in SDXL? Do you agree with this statement of diminishing returns or do you think there's still a lot to squeeze in sdxl? As the finetuner of one of the most advanced checkpoints, i'm refering to you of course, I'm curious to know what is your opinion about all of this. I have a feeling we could reach the limit soon of what SXDL is capable of. Of course, i could be wrong about my assumptions. If you don't mind asking, what's your take on all of this?

    Using Epicrealism was a good call. It really improved a lot gonzalomo 3.0 giving so much details.

    GBRX
    Author
    Jun 26, 2025· 6 reactions

    While I really appreciate all the nice compliments I do have to say that my models are checkpoint merges rather than finetunes. I'm not finetuning with my own carefully crafted dataset of images. Essentially, I'm smashing a bunch of models together and hoping for the best. However, I am trying to tweak everything to achieve the look that works best for me and I'm happy that others have a need and appreciation for a similar look.

    As far as SDXL goes, it definitely has its limits, as does everything. With my own model, I saw a big jump in quality from v1.0 to v2.0 but less of a jump from v2.0 to v3.0. I suspect that many won't see any jump from v3.0 to v4.0. So I think I'm already reaching my limit.

    That said, as someone who went exclusively Flux for about 6 months, I think SDXL is better overall in terms of flexibility and photorealism than any other image generation architecture. Flux is better at lighting and composition but SDXL is better at everything else. Some of my best images are Flux generations that have been heavily refined by SDXL.

    So, while SDXL alone is limited, it has the ability to go farther as a refiner. When an architecture comes along that can combine the best of Flux and HiDream with the best of SDXL then SDXL will die. Until then, it's still the best image gen tool as far as I'm concerned.

    ElectricDreamsJun 26, 2025

    Very interesting. Thanks for replying. I definitely see a lot of improvement between v.3.0 and v.4.0. and I know, reading the reviews, i'm not the only one. Excellent work and keep cooking!

    GBRX
    Author
    Jun 26, 2025

    @ElectricDreams Thanks!

    BuffyDuskJun 27, 2025· 3 reactions
    CivitAI

    Getting better and better!

    RakibJun 28, 2025· 2 reactions
    CivitAI

    I think v4 every image generate same face.. how to fix same face?

    GBRX
    Author
    Jun 28, 2025· 1 reaction

    I haven't noticed facial diversity being worse in v4 versus any other versions. If you don't vary the prompt and only the seed then it will maintain similar faces but that's a feature for me.

    2373972Jul 1, 2025· 4 reactions
    CivitAI

    I fell in love with this checkpoint

    JabGalJul 2, 2025· 6 reactions
    CivitAI

    Fantastic, as usual

    GBRX
    Author
    Jul 2, 2025

    Thanks!

    BkunsayJul 2, 2025· 3 reactions
    CivitAI

    This is a great model, the best SDXL I've tested so far. I'm very happy and impressed. Please keep developing it. Thank you very much.

    GBRX
    Author
    Jul 2, 2025

    Thanks for the kind comments. I'm sure I'll be tinkering with it for some time but I'm not sure I have the skills yet to make it much better. There are some advanced model merging techniques I still need to master.