Since the new “cursor-fast” model is supposed to be smarter than gpt 3.5, should we use it as the “Reranker” when doing @Codebase requests? Have you done any benchmarks in that regard?
1 Like
Since the new “cursor-fast” model is supposed to be smarter than gpt 3.5, should we use it as the “Reranker” when doing @Codebase requests? Have you done any benchmarks in that regard?