OCI Enterprise AI Models provides managed foundation-model capabilities oriented around three principal inference tasks: Chat, Embeddings, and Rerank . Chat models generate conversational or instructional responses and form the reasoning/generation foundation for many agentic applications. Embed models transform text or other supported content into numerical vector representations, enabling semantic search, recommendations, clustering, classification, and retrieval-augmented generation. Rerank models take an initial collection of retrieved candidates and reorder them according to relevance to a query, improving retrieval quality before selected context is passed to a generative model.

Oracle's current OCI Generative AI documentation explicitly identifies Chat, Embeddings, and Rerank as core Enterprise AI Model tasks. Robotics is not one of the defined Enterprise AI model categories, while clustering and classification are applications of embeddings rather than independent model categories. SQL, NoSQL, and Graph describe database technologies rather than generative-model classes.

The uploaded question source also identifies the Chat/Embed/Rerank combination as the correct selection.

Study Guide reference/topic: OCI Enterprise AI Agents — Enterprise AI Models, chat inference, embeddings, reranking, and model-supported agent workflows.

===============