Back to Arena

Redis Vector

by Redis Ltd.

System Card

OrganizationRedis Ltd.
Released2022-04
Architecturevector-rag / in-memory vector store with persistence
DetailsRedis added vector search (RediSearch module) via HNSW and FLAT indexing on high-dimensional embeddings stored as native Redis data types. Queries combine metadata filtering (with Redis query syntax) and KNN vector search in microsecond latency. Available in Redis Stack (self-hosted) and Redis Cloud (managed). Benchmarks show up to 9.5x higher QPS than pgvector.
Parameters—
Domainrag-retrievalagent-memoryepisodic-session
Open SourcePartial
WebsiteVisit
in-memorylow-latencyhybridHNSWcaching

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
59.211p
MemoryBank
no data
Cross-Session Memory
1/1
LongMemEval
58.88p
Multi-Hop QA
2/3
BABILong
no data
MultiHop-RAG
6315p
HotpotQA
60.212p
Agent Task Memory
1/1
AgentBench-Mem
63.711p
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
0/1
RAGAS
no data
Sources:Redis Vector vendor documentation; evaluated on AgentBench Memory Track (Tsinghua KEG, 2308)Redis Vector vendor documentation; evaluated on HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering (Stanford / CMU, 1809)Redis Vector vendor documentation; evaluated on LoCoMo: Long-Term Conversational Memory Benchmark (Snap Research, 2402)Redis Vector vendor documentation; evaluated on LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory (Salesforce AI Research, 2410)Redis Vector vendor documentation; evaluated on MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries (HKUST, 2401)Redis Vector vendor documentation; evaluated on LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding (Tsinghua KEG, 2308)