XAI unveils its new image generation model, Aurora. It is an autoregressive mixture of experts' networks trained to predict the next token from interleaved text and image data. The model has been trained with billions of examples from the internet enabling it to have a better and deeper understanding of the world. 

Other than generating human portraits, the model is also capable of generating text-embedded images and editing user-given images. It is available on X for public use, free of cost. The model can generate images in genres like artistic, memes, realistic & celebrities, the blog by XAI stated.  

https://t.co/mM0p2MmCa3

— xAI (@xai) December 9, 2024

Highly intense training 

The company claims that the recently launched model generates photorealistic images with tons of details that give the image a realistic rendering making it look much more real-world than a typical AI-generated image. The training provided to the model, making it interact with billions of images over the internet has helped out the developers to enable this feature. 

Varied inputs 

The model is capable of taking varied inputs and not a single type of command parallel with some specific input that the model has been trained for. It allows the model to go beyond text addition and editing. The model has native support for multimodal input, allowing it to take inspiration from or directly edit user-provided images.  

Image editing capability 

The model can also take images as input and enable the users to have better and greater control over the elements of the images enabling a render of an all-new image based on the input provided by the user. The feature is currently under development and will be available to the users on X, very soon. 

Also Read: OpenAI unveils OpenAI o1, with image upload support and a 34% performance boost

Availability 

Grok's new capabilities are now available on the X platform in select countries and will roll out to all users within a week. 

Image Gallery