Microsoft's MIT-licensed 27B decoder-only multilingual embedding model, #1 on Multilingual MTEB v2 at release.
A solid 27B-parameter dense embedding model from Microsoft. Treat the modality benchmarks above as the leading indicator of fit — composite scoring across modalities is still maturing.
Generated from this model’s benchmarks and ranking signals. Editor reviews refine it over time.
Access model weights, configuration files, and documentation.
See which devices can run this model and at what quality level.
Microsoft's flagship open-source multilingual embedding model, built on a Gemma 3 27B decoder-only backbone with last-token pooling and L2 normalization. Trained via contrastive learning across retrieval, clustering, STS, classification, bitext mining, and reranking; supports 94 languages and 32K-token contexts. Topped Multilingual MTEB v2 at release with a commercial-friendly MIT license.
View on Hugging Face