CivArchive
    Dex's Mix - v2.0
    NSFW
    Preview 94497659
    Preview 94497683
    Preview 94497714
    Preview 94497686
    Preview 94497697
    Preview 94497711
    Preview 94497694
    Preview 94497702
    Preview 94497689
    Preview 94497665
    Preview 94497661
    Preview 94497680
    Preview 94497690
    Preview 94497693
    Preview 94497691

    For the best results like the images I provided I recommend using my workflow with comfyui.

    You shouldn't need a shit ton of prompts I never had to.

    My Text 2 Image: Dex's Simple Text2Img HiRes Fix - v2.0 | Other Workflows | Civitai

    I use the easy to install app here: ComfyUI | Generate video, images, 3D, audio with AI

    Description

    Model V2 - Enhanced Quality & Style Versatility

    What's New in V2

    Major Improvements:

    • Enhanced Lighting System: Warmer, more natural lighting with improved atmospheric rendering

    • Superior Detail Generation: Automatic high-detail output without complex prompting

    • Better Style Interpretation: Fixed literal prompt interpretation issues while maintaining creative range

    • Improved Atmospheric Effects: Enhanced particle effects, subsurface scattering, and environmental depth

    Key Features

    Minimal Prompting Philosophy

    This model is designed for simplicity and consistency. No complex negative prompts or extensive quality tags needed.

    Basic Formula:

    (Score_9, Score_8_up, Score_7_up), [STYLE], Female, anthro

    Extensive Style Range

    Seamlessly handles diverse art styles with the same simple prompting approach:

    • Traditional Media: Watercolor, oil painting, pencil sketch, charcoal

    • Digital Art: Concept art, digital painting, matte painting

    • Animation Styles: Disney, Pixar/3D, anime, cartoon

    • Modern Aesthetics: Cyberpunk, vaporwave, synthwave

    • Classic Techniques: Pixel art, impressionist, art nouveau

    • Atmospheric Moods: Fantasy, gothic, cottagecore, dark academia

    Technical Strengths

    • Professional-grade lighting across all styles

    • Authentic style replication without filter-like artifacts

    • Creative flexibility - generates diverse content with minimal guidance

    Usage Examples

    Style Swapping:

    • Cyberpunk: (Score_9, Score_8_up, Score_7_up), (Cyberpunk), Female, anthro

    • Watercolor: (Score_9, Score_8_up, Score_7_up), (Watercolor), Female, anthro

    • Disney Style: (Score_9, Score_8_up, Score_7_up), (Disney Style), Female, anthro

    Add Specifics as Needed:

    • Character details: red hair, blue eyes, wolf ears

    • Poses: sitting, standing, walking

    • Settings: forest, city, indoors, fantasy castle

    • Clothing: medieval dress, modern clothes, armor

    Perfect For

    • Artists wanting consistent, high-quality results

    • Users who prefer simple, predictable prompting

    • Style experimentation and exploration

    • Professional-quality artwork without prompt engineering

    • Both creative freedom and directed generation

    Version History

    • V1: Original merge with cooler lighting, more literal style interpretation

    • V2: Enhanced lighting, improved style handling, better atmospheric rendering through strategic re-merging

    FAQ

    Comments (4)

    KaladaeAug 16, 2025
    CivitAI

    Is V2 supposed to be marked SDXL1.0? It's using PONY6 score tags.

    dexiomart555
    Author
    Aug 16, 2025· 1 reaction

    apologies i didn't notice that i'll fix it now.

    KaladaeAug 16, 2025

    dexiomart555 No problems, just wanted to make sure. Thanks for the correction.

    StreamTabulousSep 3, 2025· 1 reaction
    CivitAI

    amazing love it, gave buzz tip <3

    Checkpoint
    Pony

    Details

    Downloads
    129
    Platform
    CivitAI
    Platform Status
    Available
    Created
    8/16/2025
    Updated
    8/24/2026
    Deleted
    -

    Files

    dexsMix_v20.safetensors

    Mirrors

    CivitAI (1 mirrors)
    TensorArt (1 mirrors)

    Available On (1 platform)

    Same model published on other platforms. May have additional downloads or version variants.