multilingual-e5-large
Sovereignty and hosting
This model is catalogued for reference. It is not hosted on LLM EU yet, and no first-party endpoint serves it here. Requests to it go to the provider directly, under that provider's own terms.
Summary
multilingual-e5-large is an XLM-RoBERTa-based text embedding model with 560M parameters, published under the MIT licence and covering around 100 languages. The model card uses a maximum sequence length of 512 tokens, which is the effective limit for both indexing and querying. It has no instruction prefix for queries, so documents and queries must be embedded the same way.
Specifications
- Provider
- Microsoft (US)
- Context window
- 512
- Parameters
- 0.56B
- Modalities
- embedding
- License
- mit · licence
- Open weights
- yes
- Release date
- 30 Jun 2023
- Tasks
- rag, multilingual-support, private, cheap
Languages
Strong: en, de, fr, es, it, pt, nl, pl
Adequate: sv, da, fi, cs, ro, el, uk, ru, ar, hi, zh, ja, ko, tr
Languages
Strong
- en
- de
- fr
- es
- it
- pt
- nl
- pl
Adequate
- sv
- da
- fi
- cs
- ro
- el
- uk
- ru
- ar
- hi
- zh
- ja
- ko
- tr
Not verified: the provider has not documented this claim. — a language is listed as strong only where a source claims it; the card links the evidence where one exists.
When to use it
Strengths
- MIT licence and roughly 100 languages declared
- runs on CPU for small indexes
- long-established baseline with community tooling
Limits
- 512-token input limit
- no instruction-aware query encoding
- no retrieval benchmark has been run by LLM EU