Open-weight text-to-image model with advanced prompt adherence, anatomically accurate details, and powerful tools for inpainting, outpainting, and structural edits.