Have you seen the outputs from OpenAI's latest image model? I believe it is called "Image 2.0", is just a few weeks old and - to my untrained eyes - the outputs including text are astonishing.