What embedding dimensions does Pinecone support?

Pinecone supports arbitrary embedding dimensions up to 20000, though practical limits depend on the chosen index type and performance targets. Common dimensions match popular embedding models: 1536 for OpenAI text-embedding-3-small and ada-002, 3072 for text-embedding-3-large, 1024 for Cohere embed-v3, and 768 for many open-source sentence-transformers. Higher dimensions increase storage cost and search latency roughly linearly, so choose the smallest dimension that meets your retrieval quality target. The dimension is fixed at index creation and cannot be changed later.