BGE-M3 Embeddings (1024 dim)
Default knowledge-base embeddings - multilingual dense vectors
$1 in free credits
Overview
BGE-M3 is BAAI's multi-functional retrieval model. The checkpoint can produce dense, sparse (lexical) and multi-vector representations, works across more than 100 languages, and handles inputs from short sentences up to 8192 tokens. Knowledge bases here use its 1024-dimension dense vectors, followed by a reranker as BAAI recommends.
Strengths
- Strong multilingual dense retrieval
- More than 100 languages
- Documents up to 8192 tokens
Use cases
- General knowledge-base and RAG retrieval
- Multilingual corpora
Good to know
Used through knowledge bases or the workspace embed endpoint; the model API has no embeddings route. Dense vectors only: the checkpoint's sparse and multi-vector outputs are not served. Text only -- use qwen3-vl-embedding if images have to live in the same index. Inputs beyond 8192 tokens must be chunked.
API access
One OpenAI-compatible API for every model in the catalog, plus the CLI and the function SDK inside workflows. Sign up and try $1 of inference free, then get $10 more once you've spent $5, and these requests work:
# bge-m3 is an embedding model; the model API has no embeddings route. # A knowledge base uses it when you pick it at creation. curl https://model.graphn.ai/v1/models \ -H "Authorization: Bearer $GRAPHN_API_KEY"
graphn model get bge-m3
@function
async def embed(text: str) -> list[float]:
return await kb.embed(text)