Reka AI

AI research and product company building state-of-the-art multimodal foundation models purpose-built for understanding text, images, video, and audio natively. Founded by ex-DeepMind and Google Brain researchers, Reka develops a suite of models (Core, Flash, Edge/Spark) ranging from heavyweight multimodal reasoning to ultra-compact on-device inference. The company also offers enterprise platforms including Reka Vision for video/image understanding at scale and Reka Research for agentic AI. Backed by Nvidia, Snowflake, and DST Global with $168M raised, Reka focuses on efficient training and flexible deployment — cloud, on-prem, VPC, or fully air-gapped.