Why did we open-source our inference engine? Read the post

← Catalog

Snowflake/snowflake-arctic-embed-s

Open comparison →

Primitive: /encode · Encode · BERT

Snowflake's Arctic-embed-s

Dense

Overview

Hardware: — drives latency, throughput & cost

Size33M params
Tasks /encode
Licenseapache-2.0
Latency27 ms
Throughput99.0K tok/s
Cost$0.0022 /1M tok

Cost is approximate — computed from list GPU prices; your actual price depends on the provider you deploy SIE with.

Embedding

Output typesDense
Dimensionsdense: 384
Max sequence length512
Inputstext

Benchmarks

CQADupstackPhysicsRetrieval

scientific retrieval en

Duplicate question retrieval from StackExchange Physics

Corpus: 38,314 Queries: 1,039
Quality
ndcg at 10 0.4629
map at 10 0.4070
mrr at 10 0.4648
Performance L4 b1 c16
Corpus 66.0K tok/s
Corpus p50 25.4ms
Query 5.6K tok/s
Query p50 25.0ms
Reference →

CosQA

technology retrieval en

Code search with natural language queries

Corpus: 6,267 Queries: 500
Quality
ndcg at 10 0.3384
map at 10 0.2721
mrr at 10 0.2891
Performance L4 b1 c16
Corpus 24.9K tok/s
Corpus p50 25.6ms
Query 3.1K tok/s
Query p50 25.6ms
Reference →

FiQA2018

finance retrieval en

Financial opinion mining and question answering

Corpus: 57,599 Queries: 648
Quality
ndcg at 10 0.3972
map at 10 0.3178
mrr at 10 0.4702
Performance L4 b1 c16
Corpus 78.9K tok/s
Corpus p50 27.2ms
Query 6.1K tok/s
Query p50 24.5ms
Reference →

LegalBenchConsumerContractsQA

legal retrieval en

Question answering on consumer contracts

Corpus: 153 Queries: 396
Quality
ndcg at 10 0.7474
map at 10 0.6905
mrr at 10 0.6913
Performance L4 b1 c16
Corpus 172.1K tok/s
Corpus p50 29.8ms
Query 8.6K tok/s
Query p50 24.8ms
Reference →

NFCorpus

medical retrieval en

Biomedical literature search from NutritionFacts.org

Corpus: 3,593 Queries: 323
Quality
ndcg at 10 0.3265
map at 10 0.1209
mrr at 10 0.5400
Performance L4 b1 c16
Corpus 135.2K tok/s
Corpus p50 28.9ms
Query 2.6K tok/s
Query p50 25.4ms
Reference →

SCIDOCS

scientific retrieval en

Citation prediction, document classification, and recommendation for scientific papers

Corpus: 25,656 Queries: 1,000
Quality
ndcg at 10 0.1941
map at 10 0.1165
mrr at 10 0.3400
Performance L4 b1 c16
Corpus 92.0K tok/s
Corpus p50 26.4ms
Query 5.8K tok/s
Query p50 25.5ms
Reference →

SciFact

scientific retrieval en

Scientific claim verification using research literature

Corpus: 5,183 Queries: 300
Quality
ndcg at 10 0.6985
map at 10 0.6540
mrr at 10 0.6683
Performance L4 b1 c16
Corpus 125.3K tok/s
Corpus p50 27.4ms
Query 8.5K tok/s
Query p50 25.4ms
Reference →

StackOverflowQA

technology retrieval en

Programming question answering from Stack Overflow

Corpus: 19,931 Queries: 1,994
Quality
ndcg at 10 0.8303
map at 10 0.8014
mrr at 10 0.8014
Performance L4 b1 c16
Corpus 106.1K tok/s
Corpus p50 27.6ms
Query 131.6K tok/s
Query p50 28.5ms
Reference →

Open source inference for agents

Open-source inference for the models behind your agents. Run it yourself, or let us run it for you.

Github 2.8K

Contact us

Tell us about your use case and we'll get back to you shortly.

Apply for an inference grant

Free capacity on our hosted cluster for selected projects. Tell us what you run and we reply by email.