---
license: apache-2.0
base_model: Qwen/Qwen3-Embedding-0.6B
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- agent-skill-retrieval
---

# R3-Embedding-0.6B

The latest agent skill retrieval model at the 0.6B scale. **R3-Embedding** is the bi-encoder
(recall) stage of R3-Skill's two-stage retriever for query-conditional agent skill retrieval. It
embeds a query and every skill independently and ranks candidates by cosine similarity, paired
with [R3-Rerank-0.6B](https://huggingface.co/tencent/R3-rerank-0.6b) for reranking.

- 📄 Paper: [Skill Is Not Document: A Query-Conditional Benchmark and Two-Stage Retriever for LLM Agent Skill Routing](https://arxiv.org/abs/2606.03565)
- 💻 Code: [Tencent/R3-Skill](https://github.com/Tencent/R3-Skill)

## Usage

```python
from sentence_transformers import SentenceTransformer

model = SentenceTransformer("tencent/R3-embedding-0.6b")
query_embedding = model.encode_query("I need to compose music")
document_embeddings = model.encode_document([  # The format is "name | description | skill_md"
    "music-composer | Composes original music | Creates music for various media formats ...",
    "music-lyricist | Writes lyrics for songs | Creates lyrics for various music genres ...",
    "music-editor | Edits and mixes music tracks | Provides audio editing and mixing services ...",
])
similarities = model.similarity(query_embedding, document_embeddings)
print(similarities)
# tensor([[0.7410, 0.5510, 0.5028]])
```

## Citation

```bibtex
@inproceedings{r3skill2026,
  title  = {Skill Is Not Document: A Query-Conditional Benchmark and Two-Stage Retriever for LLM Agent Skill Routing},
  author = {Wang, Zifei and Wen, Wei and Ji, Qiang and Qiao, Ruizhi and Sun, Xing},
  year   = {2026},
  url    = {https://arxiv.org/abs/2606.03565},
}
```