ConstReadonlyada: "text-embedding-ada-002"ReadonlyLarge: "text-embedding-3-large"Most capable embedding model for both english and non-english tasks Output Dimension: 3,072
Readonlysmall: "text-embedding-3-small"Increased performance over 2nd generation ada embedding model Output Dimension: 1,536
Most capable 2nd generation embedding model, replacing 16 first generation models Output Dimension: 1,536