The new version can generate 8K, 360-degree panoramas.
Since this is only a beta version, the tool has a simplified UI and fewer features than its previous iteration, GauGAN 2, lacking the ability to turn text inputs into images. Presumably, the tool will be further improved down the line.
The newly-released AI is only a part of the new tools and programs introduced by NVIDIA during the aforementioned conference. Besides GauGAN 360, the list of new SDKs revealed by the company includes NeuralVDB, a ground-breaking update to industry standard OpenVDB, Kaolin Wisp, a Pytorch library establishing a framework for neural fields research, 3D MoMa, a new inverse rendering pipeline that allows users to quickly import a 2D object into a graphics engine, and more. You can check out the full list by clicking this link.