All models
BAAI
BGE-M3
Multilingual embeddings from BAAI
Text
Vector 1024
- Type
- Embeddings
- Context
- 8K tokens
- Russian
- high
- Price
- 0,8 ₽ / 1M tokens
- Vector size
- 1024
- Release date
- January 2024
chat/completions
import OpenAI from "openai"const client = new OpenAI({ baseURL: "https://api.yenisei.ru/v1", apiKey: process.env.YENISEI_API_KEY,})const stream = await client.chat.completions.create({ model: "bge-m3", messages: [{ role: "user", content: "Hi!" }], stream: true,})for await (const chunk of stream) { process.stdout.write(chunk.choices[0]?.delta?.content ?? "")}About the model
BGE-M3 is a multilingual embeddings model from BAAI that turns text into vector representations for semantic search and RAG. It supports many languages, long input and three modes of operation (dense, sparse and multi-vector), so you can pick the mode that fits your search pipeline. The model matches a query to the relevant documents well across languages. The resulting vectors are handy for clustering, recommendations, deduplication and classification. It handles Russian and dozens of other languages confidently.
Strengths
- Text in many languages, including Russian
- Support for long input text
- Dense, sparse and multi-vector modes
- High-quality search and RAG
- Vectors for clustering and recommendations
- Flexible for different pipelines
Where to use it
- Semantic search across documents
- RAG pipelines and question-answering systems
- Clustering and topic grouping
- Recommendation systems
- Deduplication and similarity search
- Vector-based text classification