Back to Arena

RAGFlow

by InfiniFlow

System Card

OrganizationInfiniFlow
Released2024-03
Architecturehybrid / Deep document understanding RAG
DetailsEnterprise RAG engine combining deep document layout understanding, template-based chunking, citation grounding, and agent workflows. Microservices architecture with Docker/GPU deployment.
Parameters—
Domainrag-retrievalagent-memory
Open SourceYes
WebsiteVisit
CodeRepository
enterprisedocument-layoutcitationagent-templatesopen-source

Capability Profile

Benchmark Scores

6 of 14 benchmarks
Long-Context Retrieval
1/5
RULER
no data
NIAH
no data
LooGLE
no data
LongBench
603p
∞Bench
no data
Multi-Turn Recall
1/2
LoCoMo
74.643p
MemoryBank
no data
Cross-Session Memory
1/1
LongMemEval
71.328p
Multi-Hop QA
2/3
BABILong
no data
MultiHop-RAG
73.565p
HotpotQA
76.278p
Agent Task Memory
1/1
AgentBench-Mem
70.522p
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
0/1
RAGAS
no data
Sources:RAGFlow (infiniflow/ragflow); evaluated on HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering (Stanford / CMU, 1809)RAGFlow (infiniflow/ragflow); evaluated on MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries (HKUST, 2401)RAGFlow (infiniflow/ragflow); evaluated on AgentBench Memory Track (Tsinghua KEG, 2308)RAGFlow (infiniflow/ragflow); evaluated on LoCoMo: Long-Term Conversational Memory Benchmark (Snap Research, 2402)RAGFlow (infiniflow/ragflow); evaluated on LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding (Tsinghua KEG, 2308)RAGFlow (infiniflow/ragflow); evaluated on LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory (Salesforce AI Research, 2410)