Model
Retrieval & Memory
mxbai-embed-large-v1
1Mixedbread's 2open 3state-of-the-art 4sentence 5embedding 6model
the six Ws · specification
W1
Who
mixedbread ai developed mxbai-embed-large-v1 as an open text embedding model.
W2
What
mxbai-embed-large-v1 is a BERT-large sentence embedding model producing 1024-dimensional vectors for semantic search and retrieval.
W3
Where
Weights are hosted on Hugging Face under the mixedbread-ai organization.
W4
When
Released March 2024 by mixedbread ai.
W5
Why
Built to match closed embedding APIs like OpenAI's while remaining small and self-hostable.
W6
With
Runs via sentence-transformers, Hugging Face transformers, or Transformers.js, or through mixedbread's hosted API.
W7
Watch
Open weights under Apache 2.0, self-hostable, commercial use permitted; maintained by mixedbread ai.
alternatives
works with
text-embeddingssemantic-searchsentence-transformersMatryoshkaopen-weights