AgentGL: Towards Agentic Graph Learning with LLMs via Reinforcement Learning

Yuanfu Sun; Kang Li; Dongzhe Fan; Jiajin Liu; Qiaoyu Tan

cs.CLApr 7, 2026

AgentGL: Towards Agentic Graph Learning with LLMs via Reinforcement Learning

Authors: Yuanfu Sun, Kang Li, Dongzhe Fan, Jiajin Liu, Qiaoyu Tan

Organizations: New York University Shanghai · New York University · Tsinghua University

Abstract

Large Language Models (LLMs) increasingly rely on agentic capabilities-iterative retrieval, tool use, and decision-making-to overcome the limits of static, parametric knowledge. Yet existing agentic frameworks treat external information as unstructured text and fail to leverage the topological dependencies inherent in real-world data. To bridge this gap, we introduce Agentic Graph Learning (AGL), a paradigm that reframes graph learning as an interleaved process of topology-aware navigation and LLM-based inference. Specifically, we propose AgentGL, the first reinforcement learning (RL)-driven framework for AGL. AgentGL equips an LLM agent with graph-native tools for multi-scale exploration, regulates tool usage via search-constrained thinking to balance accuracy and efficiency, and employs a graph-conditioned curriculum RL strategy to stabilize long-horizon policy learning without step-wise supervision. Across diverse Text-Attributed Graph (TAG) benchmarks and multiple LLM backbones, AgentGL substantially outperforms strong GraphLLMs and GraphRAG baselines, achieving absolute improvements of up to 17.5% in node classification and 28.4% in link prediction. These results demonstrate that AGL is a promising frontier for enabling LLMs to autonomously navigate and reason over complex relational environments. The code is publicly available at https://github.com/sunyuanfu/AgentGL.

Figures & tables

Figure 1: Method Overview. Equipped with graph-native search tools (Top-Left) for structural evidence mining, AgentGL employs a two-stage training strategy building on GCCL (Top-Right). The training progresses from Stage 1: Policy Bootstrapping (Bottom-Left), which uses shaped rewards to instill tool proficiency, to Stage 2: Mitigating Search Overuse (Bottom-Right), which optimizes the trade-off between search efficiency and reasoning accuracy.

GNN-Based Methods
Tasks	Node Classification							Link Prediction
Settings	In-Domain		Zero-shot Transfer					In-Domain		Zero-shot Transfer
Methods	OGB-Arxiv	OGB-Products	PubMed	Photo	Computers	Arxiv-23	Reddit	OGB-Arxiv	OGB-Products	PubMed	Photo	Computers	Arxiv-23	Reddit
GCN	$60.2$ ↑7.7	$58.8$ ↑10.8	$14.1$ ↑64.5	$8.0$ ↑42.8	$10.0$ ↑50.2	$1.9$ ↑63.6	$7.1$ ↑40.0	$55.5$ ↑38.1	$74.7$ ↑18.8	$51.0$ ↑24.8	$49.8$ ↑19.7	$50.5$ ↑24.1	$52.0$ ↑38.1	$53.2$ ↑34.8
RevGAT	$58.9$ ↑9.0	$56.7$ ↑12.9	$13.5$ ↑65.1	$12.1$ ↑38.8	$5.6$ ↑54.6	$1.3$ ↑64.2	$8.2$ ↑38.9	$59.0$ ↑34.6	$73.5$ ↑20.0	$55.8$ ↑20.0	$59.4$ ↑10.1	$50.7$ ↑23.9	$61.8$ ↑28.3	$61.8$ ↑26.2
SAGE	$60.1$ ↑7.8	$56.9$ ↑12.7	$11.1$ ↑67.5	$8.2$ ↑42.7	$7.4$ ↑52.8	$3.6$ ↑61.9	$3.1$ ↑44.0	$61.0$ ↑32.6	$73.4$ ↑20.1	$50.9$ ↑24.9	$51.6$ ↑17.9	$50.3$ ↑24.3	$50.5$ ↑39.6	$57.6$ ↑30.4

Table 1: Performance Comparison on Node Classification and Link Prediction benchmarks under In-Domain and Zero-shot Transfer settings. The best results are highlighted in bold. The metric used for two tasks is ACC (%). Red values (

\uparrow

) indicate the absolute gain over baselines. These denote the average improvement of the two RL variants; for GNN comparisons, they represent the average gain across 3B and 7B backbones.

		Datasets
GNSPB	MSO	OGB-Arxiv		OGB-Products		Arxiv-2023		PubMed		Amazon-Photo		Amazon-Computers		Reddit
		ACC (%)	#Search	ACC (%)	#Search	ACC (%)	#Search	ACC (%)	#Search	ACC (%)	#Search	ACC (%)	#Search	ACC (%)	#Search
✓	✗	$65.6$ ↑3.3	$4.00$ ↓18.8%	$74.1$ ↑2.9	$4.00$ ↓14.0%	$66.2$ ↑1.2	$3.98$ ↓10.5%	$80.8$ ↑1.9	$3.99$ ↓26.0%	$56.9$ ↑3.0	$3.99$ ↓20.5%	$67.2$ ↑1.4	$3.97$ ↓19.0%	$50.7$ ↑3.4	$3.97$ ↓13.5%
✗	✓	$62.3$ ↑6.6	$0.00$ ↑81.2%	$71.9$ ↑5.1	$0.00$ ↑86.0%	$61.8$ ↑5.6	$0.11$ ↑86.2%	$79.1$ ↑3.6	$0.02$ ↑73.2%	$54.3$ ↑5.6	$0.01$ ↑79.0%	$60.7$ ↑7.9	$0.01$ ↑80.0%	$48.7$ ↑5.4	$0.02$ ↑85.2%
✓	✓	$68.9$	$3.25$	$77.0$	$3.44$	$67.4$	$3.56$	$82.7$	$2.95$	$59.9$	$3.17$	$68.6$	$3.21$	$54.1$	$3.43$

Table 2: Ablation study on different RL training stages. The red (

\uparrow

) and green (

\downarrow

) denote absolute percentage gains and declines, respectively. #Search represents the average search count on the test set of each dataset (Budget=4). The percentage calculation and comparison for the number of searches are conducted based on the budget.

Panel of Figure 2 — Figure 2: Ablation study of AgentGL(7B)-GRPO on NC: Analysis of valid GNS counts and training rewards.

Model	OGB-Arxiv	Amazon-Photo	Avg.
AgentGL	68.9	59.9	64.4
w/o $r_{\textsc{cov}}$	65.2 ↑3.7	55.4 ↑4.5	60.3 ↑4.1
w/o CDR	65.9 ↑3.0	57.1 ↑2.8	61.5 ↑2.9
w/o RTT	65.4 ↑3.5	56.5 ↑3.4	61.0 ↑3.4
w/o GCCL	68.2 ↑0.7	59.3 ↑0.6	63.8 ↑0.6

Table 3: Component-wise ablations of AgentGL. Red numbers denote the absolute improvements.

Parameter	Datasets
Parameter	OGB-Arxiv	PubMed	Amazon-Computers
$\lambda_{r}=0.0$	67.1	80.6	65.7
$\lambda_{r}=0.5$	68.9	82.7	68.6
$\lambda_{r}=1.0$	66.9	80.1	66.2

Table 4: Impact of the hyperparameter

\lambda_{r}

Panel of Figure 3 — Figure 3: Ablation study of GCCL in different stages for AgentGL(7B)-GRPO on NC.

Domain	Dataset	#Nodes	#Edges	#Classes
Citation Network	OGB-Arxiv	169,343	1,166,245	40
Citation Network	PubMed	19,717	44,338	3
Citation Network	Arxiv-2023	46198	78548	40
Amazon Products	OGB-Products (subset)	54,025	74,420	47
Amazon Products	Amazon-Photo	48,362	500,939	12
Amazon Products	Amazon-Computers	87,229	721,107	10

Table 5: Dataset statistics used in this paper across three domains. #Classes correspond to the node classification label space; link prediction is treated as binary.

Appendix figures & tables12 assets

Supplementary material from the paper’s appendix.

Appendix

Model	OGB-Products		Amazon-Photo
Model	Var(NC)	Var(LP)	Var(NC)	Var(LP)
AgentGL-7B-GRPO	0.3	0.4	0.2	0.3
AgentGL-3B-GRPO	0.6	0.3	0.7	1.1
AgentGL-7B-R++	0.2	0.4	0.5	0.7
AgentGL-3B-R++	0.4	0.6	0.7	0.8

Appendix

Table 6: Variance of performance over 3 runs for node classification (NC) and link prediction (LP) on two datasets.

Acc (%)	$K=1$	$K=3$	$K=5$	$K=7$
OGB-Arxiv	65.7	68.1	68.9	68.6
Amazon-Photo	56.2	59.1	59.9	59.7

Appendix

Table 7: Sensitivity analysis of

K

value (AgentGL-7B-GRPO, Node Classification).

Panel of Figure 4 — Figure 4: A node classification (NC) case illustrating graph-native tool use. We highlight key reasoning sentences that drive the final decision.

Panel of Figure 5 — Figure 5: A link prediction (LP) case from Reddit. The model verifies a strong co-post motif by retrieving dense common 1-hop neighbors.

Panel of Figure 6 — Figure 6: NC prompt template (core). Placeholders in the prompt use {{…}}.

Panel of Figure 7 — Figure 7: Dataset-specific inserts for Arxiv (NC).

Panel of Figure 8 — Figure 8: Dataset-specific inserts for PubMed (NC).

Panel of Figure 9 — Figure 9: Dataset-specific inserts for Amazon (NC).

Panel of Figure 10 — Figure 10: Dataset-specific inserts for Reddit (NC).

Panel of Figure 11 — Figure 11: Link prediction (LP) prompt template (core). Placeholders in the prompt use {{…}}.

Panel of Figure 12 — Figure 12: Dataset-specific relation descriptions used in LP prompts.

Panel of Figure 13 — Figure 13: Per-pool search limits description used in LP prompts.

Explore similar work

GraphScout: Empowering Large Language Models with Intrinsic Exploration Ability for Agentic Graph Reasoning
Mar 2, 2026Yuchen Ying, Weiqi Jiang, Tongya Zheng +4LLM Reasoning with Graphs Agentic RAG
Zhejiang University, Hangzhou, China · Hangzhou City University, Hangzhou, China · Nanyang Technological University, Singapore, Singapore
Multi-Agent Agentic Graph Learning via Structural Signatures
Sep 9, 2026Liang Qu, Jianxin Li, Hua WangGraph Representation Learning LLM Reasoning with Graphs
Edith Cowan University, Perth, Australia. · Victoria University, Melbourne, Australia.
GRAIN: Bridging Name and Narrative Shifts in Real-World Graph Reasoning through Invariance-Rewarded Agentic RL
Aug 27, 2026Zike Yuan, Han Zhang, Jianzhi Yan +10LLM Reasoning with Graphs Language Model Robustness
Harbin Institute of Technology, Shenzhen, China · Peng Cheng Laboratory, Shenzhen, China · Xidian University, Xi’an, China

Mar 2, 2026J/K move · Enter open · S save

GraphScout: Empowering Large Language Models with Intrinsic Exploration Ability for Agentic Graph Reasoning

Yuchen Ying, Weiqi Jiang, Tongya Zheng, Yu Wang+3

Zhejiang University, Hangzhou, China · Hangzhou City University, Hangzhou, China · Nanyang Technological University, Singapore, Singapore

LLM Reasoning with Graphs Agentic RAG Graph Retrieval-Augmented Generation LLM Post-Training Synthetic Data Generation for Language Models

Open paper AskPDF arXiv

Knowledge graphs provide structured and reliable information for many real-world applications, motivating increasing interest in combining large language models (LLMs) with graph-based retrieval to improve factual grounding. Recent Graph-based Retrieval-Augmented Generation (GraphRAG) methods therefore introduce iterative interaction between LLMs and knowledge graphs to enhance reasoning capability. However, existing approaches typically depend on manually designed guidance and interact with knowledge graphs through a limited set of predefined tools, which substantially constrains graph exploration. To address these limitations, we propose GraphScout, a training-centric agentic graph reasoning framework equipped with more flexible graph exploration tools. GraphScout enables models to autonomously interact with knowledge graphs to synthesize structured training data which are then used to post-train LLMs, thereby internalizing agentic graph reasoning ability without laborious manual annotation or task curation. Extensive experiments across five knowledge-graph domains show that a small model (e.g., Qwen3-4B) augmented with GraphScout outperforms baseline methods built on leading LLMs (e.g., Qwen-Max) by an average of 16.7% while requiring significantly fewer inference tokens. Moreover, GraphScout exhibits robust cross-domain transfer performance. Our code will be made publicly available~\footnote{https://github.com/Ying-Yuchen/_GraphScout_}.

Abstract

Explore similar work

GraphScout: Empowering Large Language Models with Intrinsic Exploration Ability for Agentic Graph Reasoning

Multi-Agent Agentic Graph Learning via Structural Signatures

GRAIN: Bridging Name and Narrative Shifts in Real-World Graph Reasoning through Invariance-Rewarded Agentic RL