Deepseek

Chinese AI research lab that developed the DeepSeek family of large language models, including the open-weight V3 and R1 reasoning models, achieving frontier-level performance at significantly reduced training cost.