Jina Embeddings 5 Small
Awaiting activationSearch richer multilingual text with a compact model built specifically for retrieval tasks.
What it does
Jina Embeddings 5 Text Small is a fifth-generation text model with task-specific adapters and flexible vector dimensions. Jina recommends it as a starting point for text retrieval when stronger quality and long source context matter.
The Small variant has a 32,768-token source window and default 1,024-dimensional vectors. It is designed to represent multilingual text efficiently, making it relevant to teams evaluating searchable documents, knowledge collections and related content. embeddings.app exposes its query and passage retrieval modes.
Language coverage
Jina’s model card specifies coverage of 119+ languages. Test the customer languages and language pairs that matter to your business.
Official language coverage ↗Where it helps
Document search
Find useful passages in richer reports, articles and internal material.
Multilingual knowledge
Connect questions with related content in different languages.
Support assistants
Retrieve evidence for questions that need more than an exact keyword.
When to choose it
A Jina candidate when your retrieval needs favor quality and richer passages. Compare it with Nano when vector storage or response time is the more important constraint.
Awaiting provider activation and verified account pricing. The original long context does not remove this service’s input byte limit. Only text retrieval is exposed; quantized local deployment is not part of this hosted endpoint.
Official sources
Paraphrased from the provider’s documentation and original model information. Selection advice is our editorial guidance.
Jina AI · Embeddings and model guidance ↗Reviewed October 3, 2026