Qwen/Qwen3-Embedding-0.6B

Hugging Face
EMBEDDINGConcurrency Cost:1Model Size:0.6BQuant:BF16Ctx Length:32kPublished:Jun 3, 2025License:apache-2.0Architecture:Transformer1.0K Open Weights Warm

Qwen3-Embedding-0.6B is a 0.6 billion parameter text embedding model developed by Qwen, part of the Qwen3 Embedding series. Designed for text embedding and ranking tasks, it offers robust multilingual capabilities, long-text understanding, and supports user-defined output dimensions up to 1024. This model excels in text retrieval, classification, clustering, and bitext mining across over 100 languages, with a context length of 32,768 tokens.

Loading preview...

Qwen3-Embedding-0.6B: A Specialized Text Embedding Model

Qwen3-Embedding-0.6B is a 0.6 billion parameter model from the Qwen3 Embedding series, specifically engineered for text embedding and ranking tasks. It leverages the foundational Qwen3 architecture, inheriting strong multilingual capabilities and long-text understanding. This model is part of a comprehensive series that includes various sizes for both embedding and reranking, designed to offer flexibility and high performance.

Key Capabilities

  • Text Embedding and Ranking: Optimized for generating high-quality text embeddings and performing reranking for improved relevance.
  • Multilingual Support: Supports over 100 languages, including various programming languages, enabling robust multilingual, cross-lingual, and code retrieval.
  • Flexible Embedding Dimensions: Allows user-defined output dimensions ranging from 32 to 1024, catering to diverse application needs.
  • Instruction Awareness: Supports user-defined instructions to enhance performance for specific tasks, languages, or scenarios, with a recommended 1% to 5% performance improvement when used.
  • Long Context Window: Features a context length of 32,768 tokens for processing extensive texts.

Good For

  • Text Retrieval: Excels in tasks requiring the retrieval of relevant passages based on queries.
  • Text Classification and Clustering: Effective for organizing and categorizing text data.
  • Bitext Mining: Useful for identifying parallel texts across different languages.
  • Applications requiring efficiency: As the smallest model in its series, it balances performance with computational efficiency, making it suitable for resource-constrained environments.