Google has launched TxGemma, a series of open-source AI models designed to enhance drug development efficiency. The models, which include versions with 2 billion, 9 billion, and 27 billion parameters, are tailored for various therapeutic tasks and can understand molecular structures while providing explanations for their predictions. This initiative aims to streamline the notoriously slow and costly process of developing new drugs, making it more accessible for researchers and developers. Additionally, the Gemma 3 family of large language models has been introduced, featuring multimodal capabilities that allow for processing text, images, and video, with sizes ranging from 1 billion to 27 billion parameters, all available for free with open weights.