DigitalOcean Model Library

The DigitalOcean Model Library helps you discover, compare, and evaluate AI models from leading providers. Filter by capabilities, pricing, and deployment options, then test models in the Playground and deploy them with Serverless or Dedicated Inference.

Foundation models generate text or perform tasks like summarization, Q&A, and chat. Embedding models transform text into vectors and are primarily used for search, recommendations, and Retrieval-Augmented Generation (RAG). You typically combine both when building knowledge-based agents.

Yes — the Model Playground allows you to run prompts against any available foundation model, compare outputs side-by-side, and fine-tune parameters. It’s the fastest way to evaluate accuracy, speed, style, and cost.

Use an embedding model to vectorize your documents and store them in a Knowledge Base. When attached to an agent, the platform will automatically retrieve relevant context and feed it into your foundation model during inference—enabling accurate, domain-specific RAG workflows.