Halation Anima v1
Load in ComfyUI: put the file in models/diffusion_models, load it with UNETLoader (not the checkpoint loader), text encoder qwen_3_06b_base.safetensors via CLIPLoader, VAE qwen_image_vae.safetensors.
Settings: euler_ancestral / normal, cfg 4, 25 steps, render your final size in one pass (no hires second pass), short negative.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support