## Specs & substance

### Spec sheet

- **Developed by**: BAAI  
- **Model family**: BGE  
- **Category**: embeddings  
- **Modality**: Embedding  
- **Context window**: 8K tokens  
- **Architecture**: Dense  
- **Version**: M3  
- **License**: MIT  
- **Pricing**: $0.01 in  
- **Released**: Jan 2024  
- **Endpoint**: parasail-bge-m3

Multilingual, multi-functional embedding model supporting 100+ languages.

BGE-M3 is part of the BGE family by BAAI, categorized as a **embeddings** model. It supports a 8K tokens context window and is built on a Dense architecture.

Key strengths: multilingual, retrieval. Parasail serves BGE-M3 on a global fleet of current-gen GPUs behind a single OpenAI-compatible endpoint — with per-token pricing, no minimums, and dedicated capacity options when you need guaranteed throughput.