Two RAG systems both score 66% end-to-end accuracy on the same benchmark. Are they equally good?

It is a common trap to assume equal performance based on benchmark scores without verifying the evaluation methodology. Faithfulness against gold context is significantly easier to achieve than faithfulness against the actual context assembled by the retrieval system. Therefore, neither team has accurately measured the system that will actually ship. To determine which system is better, you must re-run the evaluation against the context that the system retrieves in production.