Imagen
Google DeepMind's text-to-image model family, which keeps improving resolution, in-image text rendering and generation speed, and is reachable through the Gemini API, Vertex AI or the Gemini app.
Imagen is Google DeepMind's text-to-image model family, and Imagen 4 pushes resolution, text rendering and generation speed as main improvements. Output spans photorealistic, illustrative and artistic styles, while better spelling makes images with words, such as posters, covers and comic panels, usable. Developers reach it through the Gemini API or Vertex AI and can fold it into a product with upscaling and editing steps.
