The Best 37 Knowledge Graph Tools in 2025

REBEL is a BART-based sequence-to-sequence model for end-to-end relation extraction, supporting over 200 different relation types.

A multilingual generative entity retrieval model based on mGENRE, optimized for historical texts, supporting 100+ languages with special adaptation for French, German, and English historical document entity linking.

KRISSBERT is a knowledge-enhanced self-supervised learning model for biomedical entity linking. It trains contextual encoders using unannotated text and domain knowledge to effectively address the diversity and ambiguity of entity names.

CODER is a knowledge-enhanced cross-language medical term embedding model, focusing on the task of medical terminology standardization.

CODER is a knowledge-infused cross-language medical terminology embedding model focused on medical terminology standardization tasks.

This model is a Text2Cypher model fine-tuned based on google/gemma-2-9b-it, capable of converting natural language questions into Cypher query statements for Neo4j graph databases.

Triplex is a Phi3-3.8B fine-tuned model by SciPhi.AI, specifically designed for building knowledge graphs from unstructured data, reducing knowledge graph creation costs by 98%.

GENRE is a sequence-to-sequence-based entity retrieval system that employs a fine-tuned BART architecture and generates unique entity names through constrained beam search technology.

Gemma 3.4B IT is a large language model based on text-to-text generation, specifically designed for converting natural language into Cypher query language.

REDFM is the multilingual version of REBEL, designed for multilingual relation extraction tasks, supporting relation triple extraction in 18 languages.

Fine-tuned based on the SecureBERT framework for extracting event cores, arguments, and their relationships from cybersecurity texts to construct knowledge graphs.

A triple extraction model fine-tuned based on the Flan-T5-XL model, used to identify subject-relation-object structures from text

Phi-3-mini-4k-instruct-graph is a fine-tuned version of Microsoft's Phi-3-mini-4k-instruct, specifically designed for entity relationship extraction from general text data, aiming to achieve comparable quality and accuracy to GPT-4 in generating entity relationship graphs.

A causal extraction model developed by the Taskload team led by Henry Leonardi, designed for automated information extraction tasks.

KnowGL is a model for generating and linking knowledge from text, capable of extracting entity-relation triples and linking them to Wikidata.

UIE is a unified information extraction framework capable of handling various tasks such as entity recognition, relation extraction, event detection, and sentiment analysis, with cross-task transfer and generalization capabilities.

A text-to-graph information extraction model based on Qwen-2.5-0.5B, jointly trained through reinforcement learning (GRPO) and supervised learning.

Gemma-2-2B-TR-Knowledge-Graph is a model fine-tuned based on gemma-2-2b-it, focusing on generating structured knowledge graphs from document content.

Phi-3-mini-128k-instruct-graph is a fine-tuned version of Microsoft's Phi-3-mini-128k-instruct, specifically designed for extracting entity relationships from general text data.

ICKG is a knowledge graph construction-specific instruction-following language model fine-tuned based on Vicuna-7B, excelling in extracting structured knowledge triples from text documents

A Text2Cypher model fine-tuned on Gemma-2-9b-it for converting natural language to Neo4j graph database query language Cypher

CGRE is a Chinese end-to-end relation extraction model based on BART, employing a generative approach to achieve entity relation extraction, demonstrating excellent performance on multiple Chinese relation extraction datasets.

AMRBART is a pre-trained semantic parser capable of converting sentences into Abstract Meaning Representation (AMR) graphs. The latest version, v2, is simpler, faster, and more powerful.

A multi-task joint model capable of simultaneously handling named entity recognition, relation extraction, entity mention detection, and coreference resolution tasks

GEBERT is a biomedical entity linking model based on PubMedBERT pre-training, enhanced with UMLS concept graph representation learning through a GAT graph encoder

SPBERT is an efficient BERT-based model pretrained on SPARQL queries, specifically designed for knowledge graph question answering tasks.

This is a model designed to extract relations from privacy policy texts, supporting the identification of data collection, storage, and transmission-related information.

FReE is a model based on DistilBERT, specifically fine-tuned for financial relationship type detection and classification tasks.

ULTRA is a foundation model for knowledge graph reasoning, capable of performing link prediction tasks on any multi-relational graph containing different entity/relation vocabularies, supporting zero-shot inference and fine-tuning.

SpanBERT is a pre-trained language model developed by Facebook Research, specifically fine-tuned on the TACRED dataset for relation extraction tasks.

This is a BERT-base model based on the KLUE benchmark, specifically designed for Korean relation extraction tasks.

mREBEL is the multilingual version of REBEL for multilingual relation extraction tasks, supporting 18 languages.

T5 model trained on the WikiKG90Mv2 dataset for tail entity prediction tasks in knowledge graphs

CODER is a knowledge-fused cross-language medical terminology embedding model for medical terminology standardization.

Jeux2BERT is a model based on the Flaubert language model and enhanced by integrating the JeuxDeMots lexical semantic network. It can capture the distributed features and associative properties of words and distinguish different relational characteristics between words or phrases.

Cuckoo is a small (300M parameters) information extraction (IE) model that mimics the next-token prediction paradigm of large language models by predicting the next token in a given context.

A relation embedding model fine-tuned on RoBERTa-large, specializing in lexical relation understanding and analogical reasoning tasks