Google Imagen
Google's family of text-to-image models, now being folded into the general Gemini model line.
Imagen is Google’s line of text-to-image models, the ones that turn a written description into a picture. It began as a research project in 2022 and became a product available through Google’s developer platforms, competing with Midjourney, DALL-E and Stable Diffusion. Images it produces carry SynthID, Google’s invisible watermark for AI-generated media.
Its interest now is mostly as a case study in how quickly these interfaces move. Google set 17 August 2026 as the earliest shutdown date for three Imagen 4 model IDs, directing developers to gemini-3.1-flash-image instead. It is not a simple rename: the dedicated generate_images() method is gone, and image generation runs through the same generate_content() call used for text. That reflects the wider industry shift toward a single multimodal model per family rather than a specialist endpoint per task, and it is a reminder that model APIs have the lifespan of a phone operating system, not a database.