How Much Reasoning Do Retrieval-Augmented Models Add beyond LLMs? A Benchmarking Framework for Multi-Hop Inference over Hybrid Knowledge

TL;DR

This paper introduces HybridRAG-Bench, a benchmarking framework designed to evaluate retrieval and multi-hop reasoning capabilities of models over hybrid knowledge sources, addressing challenges of data contamination and genuine reasoning assessment.

Contribution

The paper presents HybridRAG-Bench, a novel, flexible benchmark framework for evaluating retrieval-augmented models on multi-hop reasoning over hybrid knowledge, with contamination-aware features.

Findings

HybridRAG-Bench effectively distinguishes genuine retrieval from parametric recall.

Experiments show models improve reasoning over hybrid knowledge sources.

Framework supports domain-specific and time-aware evaluations.

Abstract

Large language models (LLMs) continue to struggle with knowledge-intensive questions that require up-to-date information and multi-hop reasoning. Augmenting LLMs with hybrid external knowledge, such as unstructured text and structured knowledge graphs, offers a promising alternative to costly continual pretraining. As such, reliable evaluation of their retrieval and reasoning capabilities becomes critical. However, many existing benchmarks increasingly overlap with LLM pretraining data, which means answers or supporting knowledge may already be encoded in model parameters, making it difficult to distinguish genuine retrieval and reasoning from parametric recall. We introduce HybridRAG-Bench, a framework for constructing benchmarks to evaluate retrieval-intensive, multi-hop reasoning over hybrid knowledge. HybridRAG-Bench automatically couples unstructured text and structured knowledge graph representations derived from recent scientific literature on arXiv, and generates knowledge-intensive question-answer pairs grounded in explicit reasoning paths. The framework supports flexible domain and time-frame selection, enabling contamination-aware and customizable evaluation as models and knowledge evolve. Experiments across three domains (artificial intelligence, governance and policy, and bioinformatics) demonstrate that HybridRAG-Bench rewards genuine retrieval and reasoning rather than parametric recall, offering a principled testbed for evaluating hybrid knowledge-augmented reasoning systems. We release our code and data at github.com/junhongmit/HybridRAG-Bench.