Skip to content
AnyRouter

BGE-M3

Embeddings
baai/bge-m3

Embed · 1024 dimensions · Jan 2024 · 2 upstreams

BGE-M3 is BAAI's open-weight multilingual text embedding model for dense, sparse, and multi-vector retrieval. It covers 100+ languages with an 8,192-token input window and returns a 1,024-dimension dense vector.

Needs your own keyTo use BGE-M3, add a key from OVHcloud or Nous Research. AnyRouter adds no fee.
max input
8K
dimensions
1024
input / 1M (list)
$0.01
routes
2

Upstreams

2
  1. 1OVHcloudYour keynot addedBilled to your key
  2. 2Nous ResearchYour keynot addedBilled to your key

Quick start

POST /api/v1/embeddings
curl
curl https://anyrouter.dev/api/v1/embeddings \
  -H "Authorization: Bearer $ANYROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "baai/bge-m3",
    "input": "The quick brown fox jumps over the lazy dog"
  }'

Details

Owner
Baai
Kind
Embeddings
Released
Jan 2024
Context
8K
Data policy
Provider retention
Endpoint
/api/v1/embeddings

Gallery

Images from the vendor
  • BGE-M3 MIRACL retrieval results (BAAI-reported)© BAAI
  • BGE-M3 long-document retrieval results (BAAI-reported)© BAAI