8B text-embedding release
According to the model card, this is the 8B model of the Qwen3 Embedding series built for text embedding and ranking tasks.
Open Source Model Profile · Qwen
Qwen3-Embedding-8B
Qwen3-Embedding-8B is a 7.57B-parameter Qwen embedding model for retrieval, classification, and clustering work. Its model card documents 100+ language support and configurable embedding dimensions.
- Publisher
- Qwen
- Task
- feature-extraction
- Model type
- qwen3
- License
- apache-2.0
- Library
- sentence-transformers
- Publication status
- Approved for indexing
Model overview
Qwen3-Embedding-8B is published by Qwen as a feature-extraction embedding model. The captured configuration identifies Qwen3ForCausalLM with model type qwen3, and Safetensors metadata reports 7,567,295,488 parameters. According to the model card, the Qwen3 Embedding series is designed for text embedding and ranking, and card data records apache-2.0.
Recorded capabilities
Multilingual retrieval scope
The model card says the series supports over 100 languages and targets text retrieval, code retrieval, classification, clustering, and bitext mining.
Configurable dimensions and instructions
According to the model card, output dimensions can be set from 32 to 4096, and user-defined instructions are supported with a reported 1% to 5% downstream gain.
Publisher-reported MTEB figure
The model card reports No.1 on the MTEB multilingual leaderboard with 70.58 as of June 5, 2025, as a publisher-reported result.
Use cases in the source record
- Text and code retrieval, classification, clustering, and bitext mining across the documented 100+ supported languages.
- Instruction-guided embedding with sentence-transformers, transformers, vLLM, or Text Embeddings Inference, following the publisher's English-instruction recommendation.
Limitations and unknowns
- No VRAM or hardware requirement was extracted from this record.
- Provider state is historical snapshot data and should be refreshed before being presented as current.
- Leaderboard and evaluation figures are publisher-reported model-card values and were not independently verified by Ethen.
Source and provenance
Source: Qwen/Qwen3-Embedding-8B
Captured: Unknown. Processed: 2026-09-07T19:34:35.999029+00:00.
Qwen3-Embedding-8B Highlights 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). This series inherits the exceptional multilingual capabilities, long-text understanding, and reasoning skills of its foundational model. The Qwen3 Embedding series represents significant advancements in multiple text embedding and ranking tasks, including text retrieval, code retrieval, t…
F001F002F003F004F005F006F007F010F011F012F014F015F016F017