Model by nvidia

cosmos3-nano-reasoner

Vision language model that excels in understanding the physical world using structured reasoning on videos or images.