Back to Arena

Nano GraphRAG

by gusye1234

System Card

Organizationgusye1234
Released2024-07
Architecturegraph-rag / Minimal GraphRAG reference (~1100 LOC)
DetailsSmall, hackable GraphRAG clone with swappable LLMs (OpenAI/Bedrock/DeepSeek/Ollama), embeddings (SBERT/Bedrock), vector backends (hnswlib/FAISS/Milvus-lite), and graph stores (NetworkX/Neo4j). Async, fully typed.
Parameters—
Domainrag-retrievalknowledge-graph
Open SourceYes
CodeRepository
minimalhackableasync~1100-loc

Capability Profile

Benchmark Scores

6 of 14 benchmarks
Long-Context Retrieval
2/5
RULER
71.352p
NIAH
no data
LooGLE
no data
LongBench
603p
∞Bench
no data
Multi-Turn Recall
0/2
LoCoMo
no data
MemoryBank
no data
Cross-Session Memory
1/1
LongMemEval
73.837p
Multi-Hop QA
2/3
BABILong
no data
MultiHop-RAG
7579p
HotpotQA
73.763p
Agent Task Memory
0/1
AgentBench-Mem
no data
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
1/1
RAGAS
70.858p
Sources:Nano GraphRAG (gusye1234/nano-graphrag); evaluated on HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering (Stanford / CMU, 1809)Nano GraphRAG (gusye1234/nano-graphrag); evaluated on MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries (HKUST, 2401)Nano GraphRAG (gusye1234/nano-graphrag); evaluated on RAGAS: Automated Evaluation of Retrieval-Augmented Generation (Exploding Gradients, 2309)Nano GraphRAG (gusye1234/nano-graphrag); evaluated on LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding (Tsinghua KEG, 2308)Nano GraphRAG (gusye1234/nano-graphrag); evaluated on LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory (Salesforce AI Research, 2410)Nano GraphRAG (gusye1234/nano-graphrag); evaluated on RULER: What's the Real Context Size of Your Long-Context Language Models (NVIDIA, 2404)