Google announced on August 20 that its open-source large model Gemma has surpassed 1 billion downloads, with over 100,000 customized variants developed to form a large open-source ecosystem known as Gemmaverse. The model is used in various fields including space image analysis, medical report standardization, cancer treatment plan design, and animal language recognition. A specialized medical version called MedGemma has been created to assist with structured medical reports and imaging analysis, and it has been used in research alongside other institutions to study cancer mechanisms. A jointly developed model called C2S-Scale 27B sorted single-cell gene data into text-like forms to help identify effective drugs for cold tumors, leading to a treatment combination that increased antigen presentation by 50% in lab tests. Though still in early research stages, the high download count for Gemma shows its value as a technical foundation and signals growing open-source use in life science research.