Imagen 4ImagesImagen 4 by Google DeepMind: photorealistic images from text. Standard, Ultra and Fast versions, right in the browser.
Imagen 4 is an image generation model from Google DeepMind, strong in photorealism, detail and prompt adherence.
The card has three versions: Standard for balance, Ultra for maximum quality, Fast for speed and lower cost. Imagen 4 works from text only; reference photos are not supported.
No, Imagen 4 works from text only. For photo editing choose GPT Image 2.5, Nano Banana or Seedream.
Standard (1 token) for most tasks, Ultra (1.5 tokens) when every detail matters, Fast (0.5 tokens) for quick drafts.
Start generating with Imagen 4 for free. Free tokens on signup.
