Back to Arena

KDB AI

by KX Systems

System Card

OrganizationKX Systems
Released2023-08
Architecturevector-rag / time-series-native vector store
DetailsKDB.AI is built on kdb+, the world's fastest time-series database, extended with HNSW vector indexing. It uniquely supports temporal queries alongside vector similarity, enabling "stateful" RAG where retrieved context can be filtered by time windows. Available as managed cloud (KDB.AI Cloud) and on-premises server edition, with FAISS and custom HNSW backends.
Parameters—
Domainrag-retrievalepisodic-session
Open SourcePartial
WebsiteVisit
time-seriestemporal-ragkdb+financial-servicesstateful

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
2/2
LoCoMo
58.29p
MemoryBank
60.15p
Cross-Session Memory
1/1
LongMemEval
56.83p
Multi-Hop QA
1/3
BABILong
no data
MultiHop-RAG
no data
HotpotQA
63.626p
Agent Task Memory
1/1
AgentBench-Mem
62.58p
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
0/1
RAGAS
no data
Sources:KDB AI vendor documentation; evaluated on AgentBench Memory Track (Tsinghua KEG, 2308)KDB AI vendor documentation; evaluated on HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering (Stanford / CMU, 1809)KDB AI vendor documentation; evaluated on LoCoMo: Long-Term Conversational Memory Benchmark (Snap Research, 2402)KDB AI vendor documentation; evaluated on LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding (Tsinghua KEG, 2308)KDB AI vendor documentation; evaluated on LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory (Salesforce AI Research, 2410)KDB AI vendor documentation; evaluated on MemoryBank: Enhancing LLMs with Long-Term Memory (Sun Yat-sen University, 2305)