AI image generators are all the rage right now, with a recent spate of upgrades to many of the available options. Ideogram is the latest synthetic image developer to join the trend this week, launching Ideogram 2.0. The new iteration of the image generator promises to outdo its predecessor and competitors with several new and improved features, as well as a new iOS app and a searchable library of the more than one billion images generated by users over the past year.
Ideogram 2.0’s text-to-image engine gives the user much more control over the styling of the AI-generated image. This includes a collection of several different styles to choose from. The Realistic style is undoubtedly the most interesting, as it produces images that closely resemble real photos. The skin, hair, and other details are much better than those of the previous Ideogram model.
The Design style, on the other hand, focuses on text accuracy within images, a notoriously difficult area for AI models to master. Ideogram 2.0 lets users generate graphic designs with long, stylized text that’s still legible. The other options are fairly self-explanatory, with 3D creating three-dimensional objects that can be rotated in real space, while Anime opts for that signature animated style, and General avoids skewing the image to a particular look.
Ideogram on the go
Ideogram 2.0 also improved its Magic Prompt and Describe tools. Magic Prompt builds on an initial prompt from a user, while Describe flips the usual setup and creates a text prompt from an image. They are now better at figuring out how to fill in details from an initially short text prompt and at explaining an image with words.
Ideogram paired its new model with the launch of its iOS app. The app will let users create and modify images directly from their mobile devices. An Android version is also in the works. In addition to the mobile app, Ideogram AI has launched a beta version of its API, allowing you to open another app or website with an AI image generator and actually interact with Ideogram’s model. It’s similar to how Microsoft uses OpenAI’s DALL-E or how X built Flux into its Grok AI chatbot. All of these are rivals to Ideogram, and while there’s no telling which, if any, will win out in the space, there’s no denying that the final image will be sharp and photorealistic, with words that anyone can read.