A compact, focused specialist that converts Japanese text into semantic embeddings. It operates with a very small context window of 514 tokens, making it suited for sentence-level tasks rather than long documents. Lightweight by design, it trades breadth for efficiency in Japanese language representation.