规模越大，BM25越赢：RAG范式没有绝对王者
检索增强生成（RAG）有四种主流范式：关键词检索、向量检索、图索引、智能体搜索。过去它们各测各的，没人知道谁在更大规模下更划算。这篇论文做了严格对照：把语料库从最小到最大拉出28个层级，跨度约450倍，问题不变，只变规模。结果没有绝对赢家，而是规模决定胜负：在最小规模下，智能体搜索（File-System Agent）领先，但它靠顺序探索，查询成本是BM25的39倍，规模一大就掉队。大约在1000万token的语料规模时，BM25反超，之后越大越赢，全规模下领先近20个百分点。向量检索一直高效但精度不够；图索引在部署前就撞上构建成本墙。结论很直接：语料越大，全局候选排序越占优，关键词检索是最稳的可扩展默认项；智能体推理适合在排序之后做精修，而不是替代排序。这不是你明天能用的技巧，但它告诉你：当你的数据量涨到千万级，别急着上复杂架构，先试试最朴素的BM25。
📄 原文摘要(英文)
Retrieval-augmented generation (RAG) spans lexical and dense retrieval, graph-based indexing, and agentic search, but these paradigms are usually evaluated on different benchmarks at one corpus size, leaving their accuracy-cost scaling unclear. To bridge this gap, we present a controlled study that varies corpus size along 28 strictly nested tiers spanning roughly 450-fold, while holding questions and a fixed bedrock of relevant and adversarial documents unchanged. Under one reader model and one judging protocol, we measure official accuracy, construction and query tokens, and latency. The results reveal a scale-dependent crossover rather than an unconditional winner. File-System Agent leads at the smallest shared tiers, but its sequential exploration costs 39 times more query tokens at the bedrock and becomes less effective as the search space grows. Around 10 million corpus tokens, BM25 overtakes it and leads at every larger shared tier, with a margin approaching 20 points at full scale. BM25 also anchors the low-cost end of the Pareto frontier without LLM-based construction. Dense retrieval remains efficient but less accurate, whereas graph-based RAG encounters construction walls before deployment scale and its scalable variants remain below BM25 at shared tiers. Overall, corpus growth increasingly favors global candidate ranking: lexical retrieval is the strongest scalable default, while agentic reasoning works best after ranked discovery rather than in place of it.