Back to Arena

Vald

by Yahoo Japan

System Card

OrganizationYahoo Japan
Released2020-02
Architecturevector-rag / cloud-native distributed ANN engine
DetailsVald is a cloud-native distributed vector search engine written in Go, designed to run on Kubernetes with automatic horizontal scaling. It uses NGT (Neighborhood Graph and Tree) as its core ANN algorithm for high-throughput approximate nearest-neighbor search. Components are independently scalable (index agent, gateway, backup manager), and it supports automatic vector index backup to object storage.
Parameters—
Domainrag-retrieval
Open SourceYes
WebsiteVisit
Kubernetescloud-nativeNGTdistributedGo

Capability Profile

Benchmark Scores

5 of 14 benchmarks
Long-Context Retrieval
2/5
RULER
69.437p
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
0/1
LongMemEval
no data
Multi-Hop QA
2/3
BABILong
no data
MultiHop-RAG
602p
HotpotQA
6011p
Agent Task Memory
0/1
AgentBench-Mem
no data
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
1/1
RAGAS
67.537p
Sources:Vald vendor documentation; evaluated on HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering (Stanford / CMU, 1809)Vald vendor documentation; evaluated on LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding (Tsinghua KEG, 2308)Vald vendor documentation; evaluated on MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries (HKUST, 2401)Vald vendor documentation; evaluated on RAGAS: Automated Evaluation of Retrieval-Augmented Generation (Exploding Gradients, 2309)Vald vendor documentation; evaluated on RULER: What's the Real Context Size of Your Long-Context Language Models (NVIDIA, 2404)