As agents operate over long horizons, their memory stores grow continuously, making retrieval critical to accessing relevant information. Many agent queries require reasoning-intensive retrieval, where the connection between query and relevant documents is implicit and requires inference to bridge. LLM-augmented pipelines address this through query expansion and candidate re-ranking, but introduce significant inference costs. We study computation allocation in reasoning-intensive retrieval pipelines using the BRIGHT benchmark and Gemini 2.5 model family. We vary model capacity, inference-time thinking, and re-ranking depth across query expansion and re-ranking stages. We find that re-ranking benefits substantially from stronger models (+7.5 NDCG@10) and deeper candidate pools (+21% from $k$=10 to 100), while query expansion shows diminishing returns beyond lightweight models (+1.1 NDCG@10 from weak to strong). Inference-time thinking provides minimal improvement at either stage. These results suggest that compute should be concentrated on re-ranking rather than distributed uniformly across pipeline stages.

翻译：随着代理在长时间跨度内执行任务，其存储的记忆会持续增长，这使得检索成为获取相关信息的关键环节。许多代理查询需要推理密集型检索，其中查询与相关文档之间的关联是隐性的，需要依据推理才能建立联系。基于大语言模型（LLM）的流水线通过查询扩展和候选重排序来解决这一问题，但会引入显著的推理成本。我们利用BRIGHT基准和Gemini 2.5模型系列，研究了推理密集型检索流水线中的计算分配问题。我们在查询扩展和重排序阶段，分别改变了模型容量、推理时思考深度以及重排序深度。研究发现，重排序阶段显著受益于更强的模型（NDCG@10提升7.5%）和更深的候选池（从k=10到k=100，提升21%），而查询扩展在超出轻量模型后收益递减（从弱模型到强模型，NDCG@10仅提升1.1%）。推理时思考在两个阶段提供的改进都微乎其微。这些结果表明，计算资源应集中于重排序阶段，而非在流水线各阶段中均匀分配。