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

Qwen/Qwen3-Embedding-0.6B

The Qwen3 Embedding model series is the latest proprietary model of the Qwen family, specifically designed for text embedding and ranking tasks. Building upon the dense foundational models of the Qwen3 series, it provides a comprehensive range of text embeddings and reranking models in various sizes (0.6B, 4B, and 8B).

Architecture
Qwen3
Parameters
596M
Tasks
Encode
Outputs
Dense
Dimensions
Dense: 1,024
Max Sequence Length
32,768 tokens
License
apache-2.0

Benchmarks

CQADupstackPhysicsRetrieval

scientific retrieval en

Duplicate question retrieval from StackExchange Physics

Corpus: 38,314 Queries: 1,039
Performance L4 b1 c16
Corpus TPS 21.2K
Corpus p50 94.7ms
Query TPS 2.7K
Query p50 58.2ms
Reference →

CosQA

technology retrieval en

Code search with natural language queries

Corpus: 6,267 Queries: 500
Performance L4 b1 c16
Corpus TPS 12.7K
Corpus p50 66.3ms
Query TPS 1.4K
Query p50 58.7ms
Reference →

FiQA2018

finance retrieval en

Financial opinion mining and question answering

Corpus: 57,599 Queries: 648
Performance L4 b1 c16
Corpus TPS 20.8K
Corpus p50 128.2ms
Query TPS 3.1K
Query p50 55.5ms
Reference →

LegalBenchConsumerContractsQA

legal retrieval en

Question answering on consumer contracts

Corpus: 153 Queries: 396
Performance L4 b1 c16
Corpus TPS 19.9K
Corpus p50 439.6ms
Query TPS 4.1K
Query p50 59.0ms
Reference →

NFCorpus

medical retrieval en

Biomedical literature search from NutritionFacts.org

Corpus: 3,593 Queries: 323
Quality
ndcg at 10 0.3689
map at 10 0.1395
mrr at 10 0.5716
Performance L4 b1 c16
Corpus TPS 21.2K
Corpus p50 240.7ms
Query TPS 1.3K
Query p50 55.9ms
Reference →

NanoFiQA2018Retrieval

finance retrieval en

Smaller subset of the FiQA financial QA dataset

Quality
ndcg at 10 0.6538
map at 10 0.5819
mrr at 10 0.7257
Performance L4 b1 c16
Corpus TPS 18.5K
Corpus p50 144.8ms
Query TPS 1.8K
Query p50 78.3ms
Reference →

SCIDOCS

scientific retrieval en

Citation prediction, document classification, and recommendation for scientific papers

Corpus: 25,656 Queries: 1,000
Performance L4 b1 c16
Corpus TPS 20.0K
Corpus p50 156.9ms
Query TPS 3.0K
Query p50 54.5ms
Reference →

SciFact

scientific retrieval en

Scientific claim verification using research literature

Corpus: 5,183 Queries: 300
Performance L4 b1 c16
Corpus TPS 21.1K
Corpus p50 218.8ms
Query TPS 3.7K
Query p50 61.3ms
Reference →

StackOverflowQA

technology retrieval en

Programming question answering from Stack Overflow

Corpus: 19,931 Queries: 1,994
Performance L4 b1 c16
Corpus TPS 20.6K
Corpus p50 172.8ms
Query TPS 18.9K
Query p50 239.7ms
Reference →

Self-hosted inference for search & document processing

Cut API costs by 50x, boost quality with 85+ SOTA models, and keep your data in your own cloud.