Awesome-Foundation-Models-on-Graphs

The is the repo for paper: Graph Foundation Models: A Comprehensive Survey

This repo will be continuously updated. Don't forget to star it and keep tuned! Please cite the paper in Citation if you find the resource helpful for your research. Thanks!

Graph Foundation Models

From Task-Specific Graph Models to General-Purpose Graph Foundation Models. (a) GFMs are pretrained on large-scale graph corpora spanning multiple domains (e.g., social, web, academic, molecular) to acquire broadly transferable representations. Through various adaptation techniques—such as fine-tuning, distillation, prompting, or zero-shot inference—they can generalize across a wide spectrum of downstream tasks, including node classification, link prediction, graph classification, and graph-to-text generation. (b) In contrast, traditional GNNs are typically trained in an end-to-end manner on a single-domain dataset for a specific task, often lacking the scalability and generalization capabilities required for open-world settings. This shift mirrors the transition observed in language and vision domains, where foundation models have redefined the standard for general-purpose intelligence.

Table of Contents

• Awesome-Foundation-Models-on-Graphs
  • Graph Foundation Models
  • Table of Contents
  • Benchmarks and Surveys
    • Benchmarks
    • Surveys
  • Universal GFMs
    • Graph Model-based Methods
    • Language Model-based Methods
    • Graph-Language Co-Training
  • Domain-Specific GFMs
    • Biology and Molecule
    • Knowledge Graph
    • Academic Graph
    • Graph Math Reasoning
    • Temporal Graph
    • Heterogeneous Graph
    • Causal Graph
    • Hyperbolic Graph
    • Document Graph
    • Graph Database
  • Task-Specific GFMs
    • Node Classification
    • Link Prediction
    • Graph Classification
    • Graph Generation
    • Question Answering
    • Recommendation
    • Anomaly Detection
  • Citation

Benchmarks and Surveys

Benchmarks

• [ICLR 25] How Do Large Language Models Understand Graph Patterns? A Benchmark for Graph Pattern Comprehension [PDF]
• [Arxiv 25.04] GraphOmni: A Comprehensive and Extendable Benchmark Framework for Large Language Models on Graph-theoretic Tasks [PDF]
• [Arxiv 25.04] OmniCellTOSG: The First Cell Text-Omic Signaling Graphs Dataset for Joint LLM and GNN Modeling [PDF] [Code]
• [Arxiv 25.02] A Comprehensive Analysis on LLM-based Node Classification Algorithms [PDF]
• [Arxiv 24.12] Benchmarking and Improving Large Vision-Language Models for Fundamental Visual Graph Understanding and Reasoning [PDF] [Code]
• [EMNLP 24] Can LLM Graph Reasoning Generalize beyond Pattern Memorization? [PDF]
• [KDD 24] LLM4DyG: Can Large Language Models Solve Spatial-Temporal Problems on Dynamic Graphs? [PDF] [Code]
• [Others 24] TreeTop: Topology-Aware Fine-Tuning for LLM Conversation Tree Understanding [PDF]
• [Others 24] FlowGPT: How Long can LLMs Trace Back and Predict the Trends of Graph Dynamics? [PDF] [Code]
• [Arxiv 24.10] Are Large-Language Models Graph Algorithmic Reasoners? [PDF] [Code]
• [Arxiv 24.07] GraphArena: Evaluating and Exploring Large Language Models on Graph Computation [PDF] [Code]
• [Arxiv 24.06] GraphFM: A Comprehensive Benchmark for Graph Foundation Model [PDF]
• [NeurIPS 23] Can Language Models Solve Graph Problems in Natural Language? [PDF]
• [Arxiv 23.10] Towards Foundational Models for Molecular Learning on Large-Scale Multi-Task Datasets [PDF]

Surveys

• [Arxiv 25.03] A Survey of Cross-domain Graph Learning: Progress and Future Directions [PDF]
• [Arxiv 25.03] Towards Graph Foundation Models: A Transferability Perspective [PDF]
• [Arxiv 25.02] Graph Foundation Models for Recommendation: A Comprehensive Survey [PDF]
• [Arxiv 25.01] Graph2text or Graph2token: A Perspective of Large Language Models for Graph Learning [PDF]
• [ICML 24] Future Directions in the Theory of Graph Machine Learning [PDF]
• [ICML 24] Graph Foundation Models are Already Here! [PDF]
• [Arxiv 24.03] A Survey on Self-Supervised Graph Foundation Models: Knowledge-Based Perspective [PDF]
• [Arxiv 23.11] Graph Prompt Learning: A Comprehensive Survey and Beyond [PDF]
• [Arxiv 23.11] A survey of graph meets large language model: Progress and future directions [PDF]
• [Arxiv 23.08] Graph Meets LLMs: Towards Large Graph Models [PDF]

Universal GFMs

Graph Model-based Methods

• [ICLR 25] Holographic Node Representations: Pre-training Task-Agnostic Node Embeddings [PDF]
• [WWW 25] RiemannGFM: Learning a Graph Foundation Model from Riemannian Geometry [PDF] [Code]
• [WSDM 25] UniGLM: Training One Unified Language Model for Text-Attributed Graph Embedding [PDF]
• [WWW 25] SAMGPT: Text-free Graph Foundation Model for Multi-domain Pre-training and Cross-domain Adaptation [PDF]
• [KDD 25] UniGraph: Learning a Unified Cross-Domain Foundation Model for Text-Attributed Graphs [PDF] [Code]
• [WWW 25] UniGraph2: Learning a Unified Embedding Space to Bind Multimodal Graphs [PDF] [Code]
• [KDD 25] Handling Feature Heterogeneity with Learnable Graph Patches [PDF] [Code]
• [KDD 25] Non-Homophilic Graph Pre-Training and Prompt Learning [PDF]
• [Arxiv 25.05] Towards Effective Federated Graph Foundation Model via Mitigating Knowledge Entanglement [PDF]
• [Arxiv 25.05] Relation-Aware Graph Foundation Model [PDF]
• [Arxiv 25.02] Boosting Graph Foundation Model from Structural Perspective [PDF] [Code]
• [EMNLP 24] OpenGraph: Towards Open Graph Foundation Models [PDF] [Code]
• [ICLR 24] One for All: Towards Training One Graph Model for All Classification Tasks [PDF] [Code]
• [NeurIPS 24] Zero-Shot Generalization of GNNs over Distinct Attribute Domains [PDF]
• [NeurIPS 24] GFT: Graph Foundation Model with Transferable Tree Vocabulary [PDF] [Code]
• [NeurIPS 24] RAGraph: A General Retrieval-Augmented Graph Learning Framework [PDF] [Code]
• [WWW 24] MultiGPrompt for Multi-Task Pre-Training and Prompting on Graphs [PDF] [Code]
• [WWW 24] Inductive Graph Alignment Prompt: Bridging the Gap between Graph Pre-training and Inductive Fine-tuning From Spectral Perspective [PDF]
• [AAAI 24] HGPROMPT: Bridging Homogeneous and Heterogeneous Graphs for Few-shot Prompt Learning [PDF] [Code]
• [Arxiv 24.12] One Model for One Graph: A New Perspective for Pretraining with Cross-domain Graphs [PDF]
• [Arxiv 24.12] Towards Graph Foundation Models: Learning Generalities Across Graphs via Task-Trees [PDF] [Code]
• [Arxiv 24.10] Towards Graph Foundation Models: The Perspective of Zero-shot Reasoning on Knowledge Graphs [PDF]
• [Arxiv 24.08] AnyGraph: Graph Foundation Model in the Wild [PDF] [Code]
• [Arxiv 24.07] Generalizing Graph Transformers Across Diverse Graphs and Tasks via Pre-Training on Industrial-Scale Data [PDF]
• [Arxiv 24.06] GraphAlign: Pretraining One Graph Neural Network on Multiple Graphs via Feature Alignment [PDF] [Code]
• [Arxiv 24.06] Cross-Domain Graph Data Scaling: A Showcase with Diffusion Models [PDF] [Code]
• [CVPR 23] Deep Graph Reprogramming [PDF]
• [NeurIPS 23] PRODIGY: Enabling In-context Learning Over Graphs [PDF] [Code]
• [KDD 23] All in One: Multi-Task Prompting for Graph Neural Networks [PDF] [Code]
• [Others 23] It's All Graph To Me: Single-Model Graph Representation Learning on Multiple Domains [PDF] [Code]
• [WWW 23] GraphPrompt: Unifying Pre-Training and Downstream Tasks for Graph Neural Networks [PDF] [Code]
• [NeurIPS 21] Transfer Learning of Graph Neural Networks with Ego-graph Information Maximization [PDF] [Code]
• [KDD 21] Adaptive Transfer Learning on Graph Neural Networks [PDF] [Code]
• [KDD 20] GCC: Graph Contrastive Coding for Graph Neural Network Pre-Training [PDF] [Code]

Language Model-based Methods

• [Arxiv 25.03] LLM as GNN: Graph Vocabulary Learning for Text-Attributed Graph Foundation Models [PDF] [Code]
• [Arxiv 25.02] Are Large Language Models In-Context Graph Learners? [PDF]
• [Arxiv 25.01] GraphICL: Unlocking Graph Learning Potential in LLMs through Structured Prompt Design [PDF]
• [ACL 24] InstructGraph: Boosting Large Language Models via Graph-centric Instruction Tuning and Preference Alignment [PDF] [Code]
• [EACL 24] Language is All a Graph Needs [PDF] [Code]
• [Others 24] GraphAgent: Exploiting Large Language Models for Interpretable Learning on Text-attributed Graphs [PDF]
• [Arxiv 24.10] LangGFM: A Large Language Model Alone Can be a Powerful Graph Foundation Model [PDF]
• [Arxiv 23.10] GraphAgent: Explicit Reasoning Agent for Graphs [PDF]
• [Arxiv 23.10] Beyond Text: A Deep Dive into Large Language Models' Ability on Understanding Graph Data [PDF]
• [Arxiv 23.07] Meta-Transformer: A Unified Framework for Multimodal Learning [PDF] [Code]
• [Arxiv 23.04] Graph-ToolFormer: To Empower LLMs with Graph Reasoning Ability via Prompt Augmented by ChatGPT [PDF] [Code]

Graph-Language Co-Training

• [Arxiv 25.03] LLM as GNN: Learning Graph Vocabulary for Foundational Graph Models [PDF] [Code]
• [ICLR 25] GOFA: A Generative One-For-All Model for Joint Graph Language Modeling [PDF] [Code]
• [ICML 24] LLaGA: Large Language and Graph Assistant [PDF] [Code]
• [SIGIR 24] GraphGPT: Graph Instruction Tuning for Large Language Models [PDF] [Code]
• [WWW 24] GraphTranslator: Aligning Graph Model to Large Language Model for Open-ended Tasks [PDF] [Code]
• [WWW 24] Can we Soft Prompt LLMs for Graph Learning Tasks? [PDF] [Code]
• [Arxiv 24.12] GraphAgent: Agentic Graph Language Assistant [PDF] [Code]
• [Arxiv 24.10] Enhance Graph Alignment for Large Language Models [PDF]
• [Arxiv 24.10] GraphCLIP: Enhancing Transferability in Graph Foundation Models for Text-Attributed Graphs [PDF] [Code]
• [Arxiv 24.10] NT-LLM: A Novel Node Tokenizer for Integrating Graph Structure into Large Language Models [PDF]
• [Arxiv 24.08] LLMs as Zero-shot Graph Learners: Alignment of GNN Representations with LLM Token Embeddings [PDF] [Code]
• [EMNLP 24] Let's Ask GNN: Empowering Large Language Model for Graph In-Context Learning [PDF]

Domain-Specific GFMs

Biology and Molecule

• [Arxiv 25.04] Integrating Single-Cell Foundation Models with Graph Neural Networks for Drug Response Prediction [PDF]
• [DAC 25] Scaling Laws of Graph Neural Networks for Atomistic Materials Modeling [PDF]
• [Arxiv 25.03] Pre-training Graph Neural Networks with Structural Fingerprints for Materials Discovery [PDF]
• [Digital Discovery 25] Hybrid-LLM-GNN: integrating large language models and graph neural networks for enhanced materials property prediction [PDF] [Code]
• [Arxiv 24.06] MolecularGPT: Open Large Language Model (LLM) for Few-Shot Molecular Property Prediction [PDF] [Code]
• [Arxiv 24.06] LLM and GNN are Complementary: Distilling LLM for Multimodal Graph Learning [PDF]
• [Arxiv 24.04] MiniMol: A Parameter-Efficient Foundation Model for Molecular Learning [PDF]
• [Arxiv 24.02] A Graph is Worth K Words: Euclideanizing Graph using Pure Transformer [PDF] [Code]
• [Arxiv 24.01] A foundation model for atomistic materials chemistry [PDF] [Code]
• [Arxiv 23.12] DPA-2: a large atomic model as a multi-task learner [PDF] [Code]
• [NeurIPS 24] On the Scalability of GNNs for Molecular Graphs [PDF] [Code]
• [NMI 24] Towards Predicting Equilibrium Distributions for Molecular Systems with Deep Learning [PDF] [Code]
• [ChemRxiv 24] Graph Transformer Foundation Model for modeling ADMET properties [PDF]
• [ICLR 24] From Molecules to Materials: Pre-training Large Generalizable Models for Atomic Property Prediction [PDF] [Code]
• [ICLR 24] BioBridge: Bridging Biomedical Foundation Models via Knowledge Graphs [PDF] [Code]
• [Others 24] GraphAgent: Exploiting Large Language Models for Interpretable Learning on Text-attributed Graphs [PDF]
• [Arxiv 23.11] InstructMol: Multi-Modal Integration for Building a Versatile and Reliable Molecular Assistant in Drug Discovery [PDF] [Code]
• [EMNLP 23] ReLM: Leveraging Language Models for Enhanced Chemical Reaction Prediction [PDF] [Code]
• [Arxiv 23.08] GIT-Mol: A Multi-modal Large Language Model for Molecular Science with Graph, Image, and Text [PDF] [Code]
• [Arxiv 23.07] MolFM: A Multimodal Molecular Foundation Model [PDF] [Code]
• [Arxiv 23.07] Can Large Language Models Empower Molecular Property Prediction? [PDF] [Code]
• [ICML 23] Enhancing Activity Prediction Models in Drug Discovery with the Ability to Understand Human Language [PDF] [Code]
• [NMI 23] Multi-modal molecule structure–text model for text-based retrieval and editing [PDF] [Code]
• [NeurIPS 23] GIMLET: A Unified Graph-Text Model for Instruction-Based Molecule Zero-Shot Learning [PDF] [Code]
• [ICLR 23] Mole-BERT: Rethinking Pre-training Graph Neural Networks for Molecules [PDF] [Code]
• [BioRxiv 22] Language models of protein sequences at the scale of evolution enable accurate structure prediction [PDF]
• [Arxiv 22.09] A Molecular Multimodal Foundation Model Associating Molecule Graphs with Natural Language [PDF] [Code]
• [EMNLP 21] Text2Mol: Cross-Modal Molecule Retrieval with Natural Language Queries [PDF] [Code]
• [NeurIPS 20] Self-Supervised Graph Transformer on Large-Scale Molecular Data [PDF]
• [Arxiv 24.08] UniMoT: Unified Molecule-Text Language Model with Discrete Token Representation [PDF] [Code]

Knowledge Graph

• [Arxiv 25.05] GraphOracle: A Foundation Model for Knowledge Graph Reasoning [PDF]
• [Arxiv 25.02] TRIX: A More Expressive Model for Zero-shot Domain Transfer in Knowledge Graphs [PDF] [Code]
• [Arxiv 25.02] How Expressive are Knowledge Graph Foundation Models? [PDF]
• [NeurIPS 24] A Foundation Model for Zero-shot Logical Query Reasoning [PDF] [Code]
• [NeurIPS 24] A Prompt-Based Knowledge Graph Foundation Model for Universal In-Context Reasoning [PDF] [Code]
• [ICLR 23] Towards Foundation Models for Knowledge Graph Reasoning [PDF] [Code]

Academic Graph

• [Arxiv 24.09] LitFM: A Retrieval Augmented Structure-aware Foundation Model For Citation Graphs [PDF]

Graph Math Reasoning

• [KDD 25] GraphTool-Instruction: Revolutionizing Graph Reasoning in LLMs through Decomposed Subtask Instruction [PDF] [Code]
• [ICLR 25] Beyond Graphs: Can Large Language Models Comprehend Hypergraphs? [PDF] [Code]
• [Arxiv 25.01] Pseudocode-Injection Magic: Enabling LLMs to Tackle Graph Computational Tasks [PDF]
• [Arxiv 24.10] A Hierarchical Language Model For Interpretable Graph Reasoning [PDF]
• [Arxiv 24.10] Graph Linearization Methods for Reasoning on Graphs with Large Language Models [PDF]
• [Arxiv 24.10] GraphTeam: Facilitating Large Language Model-based Graph Analysis via Multi-Agent Collaboration [PDF]
• [Arxiv 24.10] GCoder: Improving Large Language Model for Generalized Graph Problem Solving [PDF] [Code]
• [Arxiv 24.10] Scalable and Accurate Graph Reasoning with LLM-based Multi-Agents [PDF]
• [Arxiv 24.09] GUNDAM: Aligning Large Language Models with Graph Understanding [PDF]
• [KDD 24] GraphWiz: An Instruction-Following Language Model for Graph Problems [PDF] [Code]
• [Others 24] Can LLMs Perform Structured Graph Reasoning Tasks? [PDF] [Code]
• [Arxiv 24.03] GraphInstruct: Empowering Large Language Models with Graph Understanding and Reasoning Capability [PDF] [Code]
• [Arxiv 24.02] Let Your Graph Do the Talking: Encoding Structured Data for LLMs [PDF]
• [ICLR 24] Thought Propagation: An Analogical Approach to Complex Reasoning with Large Language Models [PDF] [Code]
• [Arxiv 23.10] GraphLLM: Boosting Graph Reasoning Ability of Large Language Model [PDF] [Code]
• [Arxiv 23.05] GPT4Graph: Can Large Language Models Understand Graph Structured Data? [PDF]
• [LoG 22] A Generalist Neural Algorithmic Learner [PDF]

Temporal Graph

• [Arxiv 25.03] Exploring the Potential of Large Language Models as Predictors in Dynamic Text-Attributed Graphs [PDF]
• [Arxiv 24.06] MiNT: Multi-Network Training for Transfer Learning on Temporal Graphs [PDF] [Code]

Heterogeneous Graph

• [AAAI 25] Bootstrapping Heterogeneous Graph Representation Learning via Large Language Models: A Generalized Approach [PDF] [Code]
• [Arxiv 25.01] HierPromptLM: A Pure PLM-based Framework for Representation Learning on Heterogeneous Text-rich Networks [PDF] [Code]
• [KDD 24] HiGPT: Heterogeneous Graph Language Model [PDF] [Code]
• [WWW 24] HetGPT: Harnessing the Power of Prompt Tuning in Pre-Trained Heterogeneous Graph Neural Networks [PDF]
• [EMNLP 23] Pretraining Language Models with Text-Attributed Heterogeneous Graphs [PDF] [Code]
• [KDD 23] Heterformer: Transformer-based Deep Node Representation Learning on Heterogeneous Text-Rich Networks [PDF] [Code]
• [KDD 23] Graph-Aware Language Model Pre-Training on a Large Graph Corpus Can Help Multiple Graph Applications [PDF]
• [KDD 22] Few-shot Heterogeneous Graph Learning via Cross-domain Knowledge Transfer [PDF]
• [KDD 21] Pre-training on Large-Scale Heterogeneous Graph [PDF]
• [NeurIPS 20] Self-supervised Auxiliary Learning with Meta-paths for Heterogeneous Graphs [PDF] [Code]

Causal Graph

• [XAI4Sci Workshop 2024] Zero-shot Causal Graph Extrapolation from Text via LLMs [PDF] [Code]

Hyperbolic Graph

• [Arxiv 25.04] HyperCore: The Core Framework for Building Hyperbolic Foundation Models with Comprehensive Modules [PDF]

Document Graph

• [ICLR 24] Harnessing Explanations: LLM-to-LM Interpreter for Enhanced Text-Attributed Graph Representation Learning [PDF] [Code]
• [Arxiv 23.11] Large Language Models as Topological Structure Enhancers for Text-Attributed Graphs [PDF] [Code]
• [Arxiv 23.10] Learning Multiplex Representations on Text-Attributed Graphs with One Language Model Encoder [PDF] [Code]
• [Arxiv 23.09] Prompt-based Node Feature Extractor for Few-shot Learning on Text-Attributed Graphs [PDF]
• [ACL 23] Patton: Language Model Pretraining on Text-Rich Networks [PDF] [Code]
• [Arxiv 23.05] ConGraT: Self-Supervised Contrastive Pretraining for Joint Graph and Text Embeddings [PDF] [Code]
• [ICLR 23] Learning on Large-scale Text-attributed Graphs via Variational Inference [PDF] [Code]

Graph Database

• [ICML 25] Griffin: Towards a Graph-Centric Relational Database Foundation Model [PDF] [Code]

Task-Specific GFMs

Node Classification

• [ICLR 25] GraphAny: A Foundation Model for Node Classification on Any Graph [PDF] [Code]
• [KDD 25] GraphLoRA: Structure-Aware Contrastive Low-Rank Adaptation for Cross-Graph Transfer Learning [PDF] [Code]
• [LoG 25] A Pure Transformer Pretraining Framework on Text-attributed Graphs [PDF]
• [WSDM 25] Training MLPs on Graphs without Supervision [PDF] [Code]
• [Arxiv 25.02] Multi-Domain Graph Foundation Models: Robust Knowledge Transfer via Topology Alignment [PDF]
• [Arxiv 25.01] Each Graph is a New Language: Graph Learning with LLMs [PDF]
• [Arxiv 24.10] Bridging Large Language Models and Graph Structure Learning Models for Robust Representation Learning [PDF]
• [NeurIPS 24] Replay-and-Forget-Free Graph Class-Incremental Learning: A Task Profiling and Prompting Approach [PDF] [Code]
• [NeurIPS 24] FUG: Feature-Universal Graph Contrastive Pre-training for Graphs with Diverse Node Features [PDF] [Code]
• [KDD 24] ZeroG: Investigating Cross-dataset Zero-shot Transferability in Graphs [PDF] [Code]
• [KDD 24] All in One and One for All: A Simple yet Effective Method towards Cross-domain Graph Pretraining [PDF] [Code]
• [WWW 24] GraphControl: Adding Conditional Control to Universal Graph Pre-trained Models for Graph Domain Transfer Learning [PDF] [Code]
• [ICLR 24] Label-free Node Classification on Graphs with Large Language Models (LLMS) [PDF] [Code]
• [Arxiv 24.01] ENGINE: Efficient Tuning and Inference for Large Language Models on Textual Graphs [PDF] [Code]
• [Arxiv 24.07] All Against Some: Efficient Integration of Large Language Models for Message Passing in Graph Neural Networks [PDF]
• [Arxiv 24.07] GraphFM: A Scalable Framework for Multi-Graph Pretraining [PDF]
• [Arxiv 24.05] LOGIN: A Large Language Model Consulted Graph Neural Network Training Framework [PDF] [Code]
• [Arxiv 24.12] Cost-Effective Label-free Node Classification with LLMs [PDF]
• [Arxiv 24.06] LangTopo: Aligning Language Descriptions of Graphs with Tokenized Topological Modeling [PDF]
• [Arxiv 24.06] Multi-View Empowered Structural Graph Wordification for Language Models [PDF] [Code]
• [Arxiv 23.10] GraphText: Graph Reasoning in Text Space [PDF] [Code]
• [SIGIR 23] Augmenting Low-Resource Text Classification with Graph-Grounded Pre-training and Prompting [PDF]
• [ICML 23] GRAFENNE: Learning on Graphs with Heterogeneous and Dynamic Feature Sets [PDF]
• [ICLR 23] Confidence-Based Feature Imputation for Graphs with Partially Known Features [PDF] [Code]
• [KDD 22] GPPT: Graph Pre-training and Prompt Tuning to Generalize Graph Neural Networks [PDF] [Code]
• [Others 21] Graph Convolutional Networks for Graphs Containing Missing Features [PDF]
• [KDD 20] GPT-GNN: Generative Pre-Training of Graph Neural Networks [PDF] [Code]

Link Prediction

• [Arxiv 25.04] Designing a reliable lateral movement detector using a graph foundation model [PDF]
• [Arxiv 25.02] How Expressive are Knowledge Graph Foundation Models? [PDF]
• [NeurIPS 24] A Prompt-Based Knowledge Graph Foundation Model for Universal In-Context Reasoning [PDF] [Code]
• [NeurIPS 24] A Foundation Model for Zero-shot Logical Query Reasoning [PDF] [Code]
• [NeurIPS 24] Universal Link Predictor By In-Context Learning on Graphs [PDF]
• [ICLR 24] Double Equivariance for Inductive Link Prediction for Both New Nodes and New Relation Types [PDF]
• [Preprint] A Multi-Task Perspective for Link Prediction with New Relation Types and Nodes [PDF]
• [ICLR 23] Towards Foundation Models for Knowledge Graph Reasoning [PDF] [Code]
• [ICLR 23] Edgeformers: Graph-Empowered Transformers for Representation Learning on Textual-Edge Networks [PDF] [Code]
• [TMLR 23] You Only Transfer What You Share: Intersection-Induced Graph Transfer Learning for Link Prediction [PDF]
• [TNNLS 23] Domain-Adaptive Graph Attention-Supervised Network for Cross-Network Edge Classification [PDF]
• [NeurIPS 21] GraphFormers: GNN-nested Transformers for Representation Learning on Textual Graph [PDF] [Code]
• [KDD 21] Cross-Network Learning with Partially Aligned Graph Convolutional Networks [PDF]

Graph Classification

• [Arxiv 24.08] UniMoT: Unified Molecule-Text Language Model with Discrete Token Representation [PDF] [Code]
• [Arxiv 24.06] LLM and GNN are Complementary: Distilling LLM for Multimodal Graph Learning [PDF]
• [Arxiv 24.03] Exploring the Potential of Large Language Models in Graph Generation [PDF]
• [Arxiv 24.02] A Graph is Worth K Words: Euclideanizing Graph using Pure Transformer [PDF] [Code]
• [Arxiv 24.01] Towards Foundation Models on Graphs: An Analysis on Cross-Dataset Transfer of Pretrained GNNs [PDF]
• [AAAI 24] G-Adapter: Towards Structure-Aware Parameter-Efficient Transfer Learning for Graph Transformer Networks [PDF]
• [AAAI 24] Fine-tuning Graph Neural Networks by Preserving Graph Generative Patterns [PDF]
• [AAAI 24] AdapterGNN: Parameter-Efficient Fine-Tuning Improves Generalization in GNNs [PDF] [Code]
• [NeurIPS 23] Universal Prompt Tuning for Graph Neural Networks [PDF] [Code]
• [NeurIPS 23] GIMLET: A Unified Graph-Text Model for Instruction-Based Molecule Zero-Shot Learning [PDF] [Code]
• [Others 23] Pretrained Language Models to Solve Graph Tasks in Natural Language [PDF]
• [ICLR 23] Mole-BERT: Rethinking Pre-training Graph Neural Networks for Molecules [PDF] [Code]
• [KDD 23] A Data-centric Framework to Endow Graph Neural Networks with Out-Of-Distribution Detection Ability [PDF] [Code]
• [IJCAI 22] Fine-Tuning Graph Neural Networks via Graph Topology induced Optimal Transport [PDF]
• [AAAI 22] Cross-Domain Few-Shot Graph Classification [PDF]
• [AAAI 21] Learning to Pre-train Graph Neural Networks [PDF] [Code]
• [NeurIPS 20] Self-Supervised Graph Transformer on Large-Scale Molecular Data [PDF]

Graph Generation

• [ICLR 25] Large Generative Graph Models [PDF] [Code]
• [NeurIPS 24] InstructG2I: Synthesizing Images from Multimodal Attributed Graphs [PDF] [Code]
• [Arxiv 24.06] Cross-Domain Graph Data Scaling: A Showcase with Diffusion Models [PDF] [Code]
• [Arxiv 24.03] Exploring the Potential of Large Language Models in Graph Generation [PDF]
• [ICLR 23] DiGress: Discrete Denoising Diffusion for Graph Generation [PDF]
• [ICML 22] Score-based Generative Modeling of Graphs via the System of Stochastic Differential Equations [PDF] [Code]
• [ICANN 18] GraphVAE: Towards Generation of Small Graphs Using Variational Autoencoders [PDF]

Question Answering

• [Arxiv 25.02] GFM-RAG: Graph Foundation Model for Retrieval Augmented Generation [PDF] [Code]
• [Arxiv 24.12] Benchmarking and Improving Large Vision-Language Models for Fundamental Visual Graph Understanding and Reasoning [PDF] [Code]
• [Arxiv 24.10] GT2Vec: Large Language Models as Multi-Modal Encoders for Text and Graph-Structured Data [PDF]
• [NeurIPS 24] G-Retriever: Retrieval-Augmented Generation for Textual Graph Understanding and Question Answering [PDF] [Code]
• [NeurIPS 24] GITA: Graph to Visual and Textual Integration for Vision-Language Graph Reasoning [PDF]
• [Arxiv 23.12] When Graph Data Meets Multimodal: A New Paradigm for Graph Understanding and Reasoning [PDF]

Recommendation

• [Arxiv 24.03] Towards Graph Foundation Models for Personalization [PDF]
• [WSDM 24] LLMRec: Large Language Models with Graph Augmentation for Recommendation [PDF] [Code]
• [WWW 24] Representation Learning with Large Language Models for Recommendation [PDF] [Code]
• [EMNLP 23] VIP5: Towards Multimodal Foundation Models for Recommendation [PDF] [Code]
• [CIKM 23] An Unified Search and Recommendation Foundation Model for Cold-Start Scenario [PDF]
• [Arxiv 21.11] Pre-training Graph Neural Network for Cross Domain Recommendation [PDF]

Anomaly Detection

• [Arxiv 25.04] GLIP-OOD: Zero-Shot Graph OOD Detection with Graph Foundation Model [PDF]
• [Arxiv 25.02] AnomalyGFM: Graph Foundation Model for Zero/Few-shot Anomaly Detection [PDF] [Code]
• [NeurIPS 24] ARC: A Generalist Graph Anomaly Detector with In-Context Learning [PDF] [Code]
• [EMNLP 24] MolCA: Molecular Graph-Language Modeling with Cross-Modal Projector and Uni-Modal Adapter [PDF] [Code]
• [ICDM 24] Towards Cross-domain Few-shot Graph Anomaly Detection [PDF]
• [Arxiv 24.10] Zero-shot Generalist Graph Anomaly Detection with Unified Neighborhood Prompts [PDF] [Code]
• [AAAI 23] Cross-Domain Graph Anomaly Detection via Anomaly-aware Contrastive Alignment [PDF] [Code]
• [TNNLS 21] Cross-Domain Graph Anomaly Detection [PDF]

Citation

@article{wang2025graph,
  title     = {Graph Foundation Models: A Comprehensive Survey},
  author    = {Wang, Zehong and Liu, Zheyuan and Ma, Tianyi and Li, Jiazheng and Zhang, Zheyuan and Fu, Xingbo and Li, Yiyang and Yuan, Zhengqing and Song, Wei and Ma, Yijun and Zeng, Qingkai and Chen, Xiusi and Zhao, Jianan and Li, Jundong and Jiang, Meng and Lio, Pietro and Chawla, Nitesh and Zhang, Chuxu and Ye, Yanfang},
  journal   = {arXiv preprint arXiv:2505.15116},
  year      = {2025}
}