Reranker models from Berlin-based Jina AI, part of Elastic since 2025: multilingual, with long context and a multimodal variant.

jina-embeddings-v5-text overview and key facts (October 2026)
Multilingual embedding models from Berlin-based Jina AI; the small version has 677 million parameters and 32,768 tokens of input.
What is jina-embeddings-v5-text?
jina-embeddings-v5-text is the fifth generation of multilingual embedding models from Jina AI in Berlin, now part of Elastic, offered in a small and a nano version for search, matching and classification.
The small model was released on February 18, 2026 with 677 million parameters, is built on Qwen3-0.6B-Base and produces 1,024-dimensional vectors that can be shortened down to 32 dimensions. It supports retrieval, text matching, clustering and classification with inputs of up to 32,768 tokens. A nano version with 239 million parameters targets constrained environments. The weights are under CC BY-NC 4.0; Elastic offers a hosted version through its inference service.
jina-embeddings-v5-text AI features and use cases
- Indexes long documents in one pieceInputs of up to 32,768 tokens reduce the need to split contracts or manuals into small chunks. Search results keep more context.
- Saves storage with shorter vectorsMatryoshka dimensions let teams store shorter vectors with limited quality loss. Large archives need less vector database capacity.
- Groups and classifies textDedicated task modes also cover clustering and classification. Service teams can sort incoming messages by topic.
What sets jina-embeddings-v5-text apart
- Developed in Berlin by Jina AI, now owned by Elastic.
- Non-commercial license for the open weights; commercial use via license or the hosted Elastic service.
Where jina-embeddings-v5-text is strongest
Jina AI GmbH is based in Berlin and has been part of Elastic since October 2025.
Questions to ask before you buy jina-embeddings-v5-text
The open weights are licensed CC BY-NC 4.0, so business use needs a commercial license from Jina AI or the hosted Elastic service. If you self-host under a license, your team runs and updates the model. Re-embedding is needed if you switch models later.
Using jina-embeddings-v5-text in the EU: GDPR and AI Act
- Data locationNot published by the vendor.
- Data processing agreementNot published by the vendor: ask for one (Art. 28 GDPR).
- EU AI ActCheck whether the AI prepares decisions about people. If not, the high-risk rules usually do not apply. The high-risk obligations apply from December 2, 2027 (Digital Omnibus).
jina-embeddings-v5-text reviews
No reviews yet. Be the first to share how your team uses it.
Every review shows its verification level. We read each one before it goes live. How we check reviews
Questions about jina-embeddings-v5-text
Can I use jina-embeddings-v5 commercially?
Not under the open license, which is CC BY-NC 4.0. Jina AI asks commercial users to contact sales.
Who owns Jina AI?
Elastic N.V. acquired Jina AI GmbH in October 2025.
How long can inputs be?
Up to 32,768 tokens for the small model.
jina-embeddings-v5-text update historyno updates yet
No changes recorded for jina-embeddings-v5-text yet.
Changes the vendor made and confirmed user reports, each checked against the vendor's own pages. We review every tool monthly.
Alternatives to jina-embeddings-v5-text
Tools in the same category (Embedding models) with the most overlapping markets.
Open reranker models from Mixedbread under Apache 2.0 for more than 100 languages and code, also available through the vendor's search API.
IBM's open enterprise models in 3B, 8B and 30B sizes, under Apache 2.0 and with ISO 42001 certification.
Open multilingual reranker model from the Beijing Academy of Artificial Intelligence under Apache 2.0, light enough for your own servers.



