AI coding glossary
Embeddings
Also known as: text embeddings, vector embeddings
In one sentence
High-dimensional numeric representations of text, code, or images that capture semantic meaning, the input to vector search, RAG, and clustering.
Full definition
Embeddings are dense numeric vectors (typically 256-4096 dimensions) produced by an embedding model, semantically similar inputs end up near each other in the vector space. They're the foundation of vector search, RAG retrieval, clustering, dedup, and recommendation systems. In 2026, common embedding models include OpenAI's text-embedding-3, Voyage's embed-3, Cohere's embed-v3, and open-source contenders like nomic-embed-text and BAAI's bge-large. Code-specific embedding models (CodeSage, code-embedding-3) outperform general-purpose embeddings on programming-language similarity tasks.