A team plans to move from one embedding model to another because the newer one scores better on a public benchmark. Their retrieval system holds four million embedded passages. What must be accounted for?

• ANothing beyond the change in per call pricing, since embeddings are a standard format.
• BEvery passage must be embedded again, because vectors from different models are not comparable.
• COnly the passages added after the switch need embedding with the new model.
• DThe vector database must be migrated to a different AWS service.