An embeddingA list of numbers representing a piece of text's meaning, so that similar meanings end up numerically close and can be searched.Full definition → model turns content into numerical representations that a search system compares to find related material. OpenRouter'sselection guide checks whether models accept requests and return the expected representations, but explicitly separates those compatibility checks from retrieval quality. To judge search performance, it recommends using real questions with labeled relevant passages and holding the documents and grading method constant across models. A successful API response establishes that the integration works; finding the passages users need requires a different test.
Terms in this piece · Glossary
embedding — A list of numbers representing a piece of text's meaning, so that similar meanings end up numerically close and can be searched.
RAG — Retrieval-augmented generation — fetching relevant documents first and pasting them into the model's context so it answers from your data instead of memory.
