Back to Arena

Marqo

by Marqo Pty Ltd

System Card

OrganizationMarqo Pty Ltd
Released2022-08
Architecturevector-rag / tensor search with integrated embedding
DetailsMarqo tightly couples embedding generation and vector indexing into one system, so documents are automatically tensorized on ingest via configurable model selection. It supports multimodal text+image queries, filtering via JSON attributes, and hybrid BM25+vector scoring. Backed by OpenSearch internals, it is offered as both open-source self-hosted and a managed cloud service.
Parameters—
Domainrag-retrieval
Open SourcePartial
WebsiteVisit
multimodale-commerce-searchtensor-searchhybridmanaged-cloud

Capability Profile

Benchmark Scores

5 of 14 benchmarks
Long-Context Retrieval
2/5
RULER
67.618p
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
64.320p
HotpotQA
60.413p
Agent Task Memory
0/1
AgentBench-Mem
no data
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
1/1
RAGAS
6843p
Sources:Marqo vendor documentation; evaluated on HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering (Stanford / CMU, 1809)Marqo vendor documentation; evaluated on LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding (Tsinghua KEG, 2308)Marqo vendor documentation; evaluated on MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries (HKUST, 2401)Marqo vendor documentation; evaluated on RAGAS: Automated Evaluation of Retrieval-Augmented Generation (Exploding Gradients, 2309)Marqo vendor documentation; evaluated on RULER: What's the Real Context Size of Your Long-Context Language Models (NVIDIA, 2404)