GraphWalk: Enabling Reasoning in Large Language Models through Tool-Based Graph Navigation

TL;DR

GraphWalk introduces a tool-based, training-free framework enabling large language models to perform multi-hop reasoning over large knowledge graphs, significantly improving performance on complex graph navigation tasks.

Contribution

It provides a universal, minimal set of graph operations for LLMs to reason over large graphs without task-specific training or domain knowledge.

Findings

GraphWalk enables LLMs to solve maze traversal problems that non-reasoning models cannot.

It achieves substantial performance gains over in-context baselines across various models.

Performance improvements are more significant with larger models where in-context methods fail.

Abstract

The use of knowledge graphs for grounding agents in real-world Q&A applications has become increasingly common. Answering complex queries often requires multi-hop reasoning and the ability to navigate vast relational structures. Standard approaches rely on prompting techniques that steer large language models to reason over raw graph context, or retrieval-augmented generation pipelines where relevant subgraphs are injected into the context. These, however, face severe limitations with enterprise-scale KGs that cannot fit in even the largest context windows available today. We present GraphWalk, a problem-agnostic, training-free, tool-based framework that allows off-the-shelf LLMs to reason through sequential graph navigation, dramatically increasing performance across different tasks. Unlike task-specific agent frameworks that encode domain knowledge into specialized tools, GraphWalk equips the LLM with a minimal set of orthogonal graph operations sufficient to traverse any graph structure. We evaluate whether models equipped with GraphWalk can compose these operations into correct multi-step reasoning chains, where each tool call represents a verifiable step creating a transparent execution trace. We first demonstrate our approach on maze traversal, a problem non-reasoning models are completely unable to solve, then present results on graphs resembling real-world enterprise knowledge graphs. To isolate structural reasoning from world knowledge, we evaluate on entirely synthetic graphs with random, non-semantic labels. Our benchmark spans 12 query templates from basic retrieval to compound first-order logic queries. Results show that tool-based traversal yields substantial and consistent gains over in-context baselines across all model families tested, with gains becoming more pronounced as scale increases, precisely where in-context approaches fail catastrophically.