GTA VI LoRA trained on the 72 images provided by Rockstar after the release of the second trailer in May 2025.

The trigger word is " gtaillvi ", better way to formulate it based on training is: " gtaillvi style of [insert your prompt here] ".

For instance: " gtaillvi style of two elderly men sitting in lounge chairs in a backyard garden with a small swimming pool. They are shirtless, wearing colorful shorts, sandals, and sunglasses, each holding a cold beer. A round plastic table with more bottles and snacks is beside them. One of them has a tattooed arm and a gold chain, the other is laughing with a missing tooth. The garden is surrounded by palm trees and a pink wall, with a flamingo float drifting in the pool. Warm sunset light, relaxed atmosphere, cinematic framing."

Had the better results with CFG between 2.5 and 3, especially when keeping the scenes simple and not too visually cluttered.