Back to Arena

Couchbase Vector

by Couchbase Inc.

System Card

OrganizationCouchbase Inc.
Released2024-05
Architecturevector-rag / multi-model NoSQL with vector index
DetailsCouchbase added HNSW-based vector search to its multi-model NoSQL platform in May 2024 (preview February 2024). Queries combine vector similarity with N1QL (SQL++) predicates over JSON documents, enabling hybrid semantic+structured search in one query. Available on Couchbase Capella (managed cloud) and Couchbase Mobile (edge devices), supporting on-device AI inference.
Parameters—
Domainrag-retrievalagent-memory
Open SourcePartial
WebsiteVisit
NoSQLN1QLedge-AIhybridmulti-model

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
6012p
MemoryBank
no data
Cross-Session Memory
1/1
LongMemEval
58.26p
Multi-Hop QA
2/3
BABILong
no data
MultiHop-RAG
65.526p
HotpotQA
62.220p
Agent Task Memory
1/1
AgentBench-Mem
60.26p
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
0/1
RAGAS
no data
Sources:Couchbase Vector vendor documentation; evaluated on HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering (Stanford / CMU, 1809)Couchbase Vector vendor documentation; evaluated on MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries (HKUST, 2401)Couchbase Vector vendor documentation; evaluated on AgentBench Memory Track (Tsinghua KEG, 2308)Couchbase Vector vendor documentation; evaluated on LoCoMo: Long-Term Conversational Memory Benchmark (Snap Research, 2402)Couchbase Vector vendor documentation; evaluated on LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding (Tsinghua KEG, 2308)Couchbase Vector vendor documentation; evaluated on LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory (Salesforce AI Research, 2410)