CivArchive
    GTA6 Cinematic Style - v1.0
    NSFW
    Preview 149723607
    Preview 149723608
    Preview 149723610
    Preview 149723609
    Preview 149723612
    Preview 149723611
    Preview 149723613
    Preview 149723614
    Preview 149723615

    This style LoRA reproduces the look of the GTA VI trailer and promo art: glossy, highly detailed cinematic 3D renders with shallow depth of field, strong rim lighting, and saturated color grading. It ranges from hot pink and violet neon nights to harsh Florida midday sun and hazy golden-hour light.

    Base model: Moss Image (trained on the Moss Image base model)

    Trigger word: gt6

    What it was trained on:

    28 widescreen images (1920×1080 and 3840×2160) with detailed natural-language captions. The subjects cover the range of the game's world:

    - Characters: close-ups and medium shots of tattooed, stylish characters with hoop earrings, gold chains, tank tops, bandanas and sunglasses, often holding pistols or rifles

    - Crime and action: car shootouts, a mall heist on dirt bikes, an armored-truck robbery, standoffs, police cars, helicopter searchlights

    - Nightlife: nightclubs and VIP sections, a waterfront dive bar, a backstage dressing room, cash and champagne

    - Places: a neon gas station and tire shop at twilight, waterfront mansions and yachts, car meets on palm-lined plazas, inside cars at night

    - Wilderness: misty swamps with alligators and airboats, a rural campfire camp, a golden hillside at sunset, an underwater coral reef

    How to prompt it:

    - Start with gt6, and then write a full sentence describing the scene. The captions are long and descriptive, so detailed prompts work better than tag lists.

    - Name the lighting and colors, as the captions do: "saturated violet and electric blue light", "hot pink and cyan lighting", "harsh midday sunlight", "golden hour".

    - Ending the prompt the way the training captions end reinforces the style: "The style is a glossy, highly detailed cinematic 3D render with shallow depth of field, conveying a tense mood."

    - Widescreen sizes such as 16:9 match the training data best.

    Description

    LORA
    MOSS Image

    Details

    Downloads
    70
    Platform
    CivitAI
    Platform Status
    Available
    Created
    11/21/2026
    Updated
    11/22/2026
    Deleted
    -

    Files