AI Models Explained – SDXL, 1.5, Flex, Rocview & More

 

AI Models Explained – SDXL, 1.5, Flex, Rocview & More

AI image generation combines powerful models like SDXL and Juggernaut with style modules (LoRAs) to create unique images from text. While the AI ​​model defines the basic style and anatomy, a LoRA adds specific flair such as Art Deco, realism, or futurism. Together, they form a flexible system that allows artists, designers, and creators to build visual scenes ranging from realistic to surreal. This technique is widely applied in digital art, game design, and concept development.

 Difference between SDXL and 1.5


👉 SDXL is ideal for cinematic images, realism, and complex compositions .
👉 SD 1.5 is fast, lightweight, and good for anime, pastels, and simple illustrations .

 Key SDXL models (rating & style)

1. Juggernaut XL v9

  • Rating : ★★★★★
  • Style : Realistic, anatomically strong, good for portraits and fashion
  • LoRA combinations : ArtDecoStyleSDXLTouch_of_Realism,add-detail-xl
  • Steps : 80–100
  • CFG : 7.5–8.5
  • Special : Strong in lighting and skin texture

2. Dreamshaper XL v2 Turbo

  • Rating : ★★★★☆
  • Style : Balanced between realism and flair, good for fantasy and deco
  • LoRA combinations : Arcane_Decopunk_Fantasy,Super_Eye_Detailer
  • Steps : 60–80
  • CFG : 7.0–8.0
  • Special : Fast generation, good color control

3. cyberrealisticXL_v70

  • Rating : ★★★★☆
  • Style : Hyperrealistic, futuristic, good for sci-fi and tech scenes
  • LoRA combinations : Cyberpunk_Techno_Mages_SDXL,add-detail-xl
  • Steps : 100–120
  • CFG : 8.0–9.0
  • Special : Strong in metal, neon, and digital atmosphere

4. RealVisXL5 (Architecture)

  • Rating : ★★★★☆
  • Style : Architecture, symmetry, composition — perfect for Deco and Rococo
  • LoRA combinations : ArtDecoStyleSDXLInspiredbyMucha,Hands
  • Steps : 80–100
  • CFG : 7.5–8.5
  • Special : Strong in structure and geometry

🧬 Other models: Flex & Rocview

Flex (SDXL Flex)

  • Style : Dynamic, good for rapid variation and test generation
  • Special : Flexible prompt interpretation, good for experimentation
  • LoRA's : Works with everything, but especially good with add-detail-xlandHands

Rocview

  • Style : Rococo, Baroque, ornate and ornamental
  • Special : Ideal for historical flair and decorative scenes
  • LoRA's : InspiredbyMuchaArtDecoStyleSDXL,Lucid_Verdant_Dream

 Main 4 Generators

  1. ComfyUI – Full control, modular, perfect for advanced users
  2. InvokeAI – User-friendly, good for fast workflows
  3. Automatic1111 (WebUI) – Most used, many plugins and LoRA support
  4. Fooocus – Simple, fast, ideal for beginners and quick test images

 Settings for optimal balance


 Upscaling – What you need to know (increasing the resolution of the captured image)

  • Goal : Increase resolution without losing detail
  • Tools : ESRGAN4x-UltraSharpRealESRGAN,Latent Upscale
  • Usage : After generation, especially with 1024×1024 images
  • Tip : Don't upscale until your composition and details are correct — otherwise you'll magnify mistakes

 Extra features per model


AI Models by Theme & Style

-Juggernaut XL v9

  • Stands out in : Animals, buildings, portraits
  • Style : Realistic, anatomically strong
  • LoRA's : Touch_of_Realismadd-detail-xl,Hands
  • Steps : 80–100
  • CFG : 7.5–8.5
  • Special : Strong in lighting, skin texture, facial expression

-Dreamshaper XL v2 Turbo

  • Stands out in : Fantasy, animals, goddesses
  • Style : Stylized realism, magical
  • LoRA's : Arcane_Decopunk_Fantasy,Lucid_Verdant_Dream
  • Steps : 60–80
  • CFG : 7.0–8.0
  • Special : Fast flair, good color control

cyberrealisticXL_v70

  • Excels in : Cars, planes, sci-fi
  • Style : Futuristic, hyperrealistic
  • LoRA's : Cyberpunk_Techno_Mages_SDXL,add-detail-xl
  • Steps : 100–120
  • CFG : 8.0–9.0
  • Special : Neon, metal, tech vibe, digital energy

-RealVisXL5 (Architecture)

  • Excels in : Buildings, structures, composition
  • Style : Architectural, geometric
  • LoRA's : ArtDecoStyleSDXLInspiredbyMucha,Hands
  • Steps : 80–100
  • CFG : 7.5–8.5
  • Special : Strong in symmetry, Deco structure

-matureRitual_v13

  • Stands out in : Fantasy, rituals, temples
  • Style : Mystical, baroque, ceremonial
  • LoRA's : Arcane_Decopunk_Fantasy,Lucid_Verdant_Dream
  • Steps : 80–100
  • CFG : 7.5–8.5
  • Special : Atmospheric, divine compositions

-animePastelDream_softBakedVae

  • Stands out in : Animals, vehicles, illustrations
  • Style : Anime, pastel, soft
  • LoRA's : EnvyCuteFairyXL01,Alien Style
  • Steps : 40–60
  • CFG : 6.5–7.5
  • Special : Soft, colorful, ideal for cute images

CyberRealistic v4.1

  • Excels in : Aircraft, technology, interfaces
  • Style : Sci-fi realism
  • LoRA's : Cyberpunk_Techno_Mages_SDXL,add-detail-xl
  • Steps : 80–100
  • CFG : 8.0–9.0
  • Special : Strong in digital structures and tech environments

-Deliberate v5

  • Stands out in : Buildings, animals, portraits
  • Style : Neutral realism
  • LoRA's : Touch_of_Realism,add-detail-xl
  • Steps : 60–80
  • CFG : 7.0–8.0
  • Special : Widely applicable, stable and reliable

-ReV Animated

  • Stands out in : Cars, action, fantasy
  • Style : Animated, dynamic
  • LoRA's : Alien Style,Cyberpunk_Techno_Mages_SDXL
  • Steps : 60–80
  • CFG : 7.5–8.5
  • Special : Strong in movement, action and visual energy

-------------------------------------------------------------------------------------------------------------------
an example 

prompt: A regal woman seated on a symmetrical Art Deco throne, wearing an emerald green gown with gold geometric patterns. Her eyes are vivid and expressive, framed by dramatic eyeliner. The background features fan-shaped columns, radiant teal neon lights, and a sunburst halo behind her. Her hands rest gracefully on the throne, adorned with cascading gold chains. The atmosphere blends 1930s elegance with futuristic mysticism.


because it is a temple you would choose mature ritual v 13 v.      but

matureRitual_v13 does indeed have a somewhat cartoonish flair , especially in the faces and clothing textures. This is due to the way the model visualizes atmosphere and ritual : she often opts for stylized shapes , bright colors , and symbolic compositions reminiscent of illustration or fantasy art.

 Why it can look cartoonish:

  • Faces are often smooth and stylized , with less skin texture
  • Clothing and accessories have a decorative, almost drawn appearance
  • Backgrounds are often rich in patterns, but sometimes lack depth or realistic lighting

What you can do about it

1. Combine with realistic LoRAs

For example:

  • Touch_of_Realism(0.3–0.4) → adds skin texture and anatomy
  • add-detail-xl→ enhances texture and depth

2. Use a more realistic model

For example:

  • Juggernaut XL v9 → retains ritualistic atmosphere but makes it more human
  • Dreamshaper XL v2 → adds flair without overstyling

3. Adjust your prompt

Use terms like:

  • “natural skin texture,” “soft lighting,” “realistic anatomy”
    And avoid:
  • “stylized,” “illustration,” “fantasy portrait” if you want realism


- The Breachline – Start & Overview



This page is part of the Noverra series, which combines stylistic shifts, rhythm, and digital image structures. Each image carries a unique codex and is aligned with a larger visual system in which form, color, and composition are given meaning. The content is rhythmically placed within the Tussen Klok en Klepel blog universe and aligns with the classification chart and cyclical structure of the series. For more context and in-depth information, visit the Noverra overview page.


Comments

Populairste blogs

A new way of thinking about digital art The Noverra-Digital Art style

Reality-tv breekt mensen, en wij kijken toe

Voor wie denkt, lacht, en af en toe gewoon even niks snapt