SnowPro Specialty Gen AI GES-C01

The EMBED_TEXT_768 function in Snowflake Cortex produces a 768-dimensional embedding of input text and returns it as a VECTOR(FLOAT, 768) value. Snowflake's VECTOR data type is purpose-built to store fixed-length arrays of numeric (float or int) values such as embeddings, enabling efficient use with vector similarity functions like VECTOR_COSINE_SIMILARITY and VECTOR_L2_DISTANCE for semantic search and retrieval-augmented generation workloads.