Back to Arena

Nomic Atlas

by Nomic AI Inc.

System Card

OrganizationNomic AI Inc.
Released2022-10
Architectureexternal-memory-network / embedding visualization and retrieval platform
DetailsNomic Atlas is an interactive data map platform for exploring, organizing, and searching large embedding datasets. It provides nomic-embed-text and nomic-embed-vision models (fully open weights on HuggingFace) that run locally or via API. Atlas visualizes millions of embeddings via UMAP/t-SNE interactively in a browser. Also maintains GPT4All, an open-source local LLM runtime.
Parameters—
Domainrag-retrievalpersonalization
Open SourceYes
WebsiteVisit
open-weightsvisualizationlocal-inferenceUMAPGPT4All

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
69.324p
MemoryBank
59.63p
Cross-Session Memory
1/1
LongMemEval
64.620p
Multi-Hop QA
2/3
BABILong
no data
MultiHop-RAG
69.735p
HotpotQA
6633p
Agent Task Memory
0/1
AgentBench-Mem
no data
Personalization
0/1
PerLTQA
no data
Factuality / Grounding
0/1
RAGAS
no data
Sources:Nomic Atlas vendor documentation; evaluated on HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering (Stanford / CMU, 1809)Nomic Atlas vendor documentation; evaluated on LoCoMo: Long-Term Conversational Memory Benchmark (Snap Research, 2402)Nomic Atlas vendor documentation; evaluated on LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding (Tsinghua KEG, 2308)Nomic Atlas vendor documentation; evaluated on LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory (Salesforce AI Research, 2410)Nomic Atlas vendor documentation; evaluated on MemoryBank: Enhancing LLMs with Long-Term Memory (Sun Yat-sen University, 2305)Nomic Atlas vendor documentation; evaluated on MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries (HKUST, 2401)