Google's Imagen 4 flagship model. This is best suited for high-quality text-to-image generation tasks where photorealism, coherence with prompts, and complex scene composition are important. It is one of Google DeepMind’s most advanced diffusion-based image generation models.