Rethinking Agentic Search with Pi-Serini: Is Lexical Retrieval Sufficient?

TL;DR

This paper demonstrates that a well-tuned lexical retriever paired with capable LLMs can effectively support deep research tasks, challenging the necessity of dense retrieval methods.

Contribution

It introduces Pi-Serini, a search agent combining lexical retrieval with browsing and reading tools, and shows its effectiveness with large language models like GPT-5.5.

Findings

BM25 tuning significantly improves answer accuracy and evidence recall.

Deeper retrieval enhances evidence recall substantially.

Lexical retrieval with proper configuration can outperform dense retrievers in research tasks.

Abstract

Does a lexical retriever suffice as large language models (LLMs) become more capable in an agentic loop? This question naturally arises when building deep research systems. We revisit it by pairing BM25 with frontier LLMs that have better reasoning and tool-use abilities. To support researchers asking the same question, we introduce Pi-Serini, a search agent equipped with three tools for retrieving, browsing, and reading documents. Our results show that, on BrowseComp-Plus, a well-configured lexical retriever with sufficient retrieval depth can support effective deep research when paired with more capable LLMs. Specifically, Pi-Serini with gpt-5.5 achieves 83.1% answer accuracy and 94.7% surfaced evidence recall, outperforming released search agents that use dense retrievers. Controlled ablations further show that BM25 tuning improves answer accuracy by 18.0% and surfaced evidence recall by 11.1% over the default BM25 setting, while increasing retrieval depth further improves surfaced evidence recall by 25.3% over the shallow-retrieval setting. Source code is available at https://github.com/justram/pi-serini.