Jina Embeddings 5 Nano
Awaiting activationCompact multilingual search vectors for applications that prioritize efficiency.
What it does
Jina Embeddings 5 Text Nano is the lighter text option in the fifth-generation Jina family. It uses task-specific adapters and flexible output dimensions, with 768-dimensional vectors by default.
Jina documents support for 15 major languages for Nano, a narrower coverage than Small’s broader multilingual foundation, and an 8,192-token source window. It is designed for teams evaluating compact representations and lighter retrieval workloads. embeddings.app exposes its text query and passage modes.
Language coverage
Jina describes Nano as multilingual, but this directory does not infer named language coverage from its EuroBERT base model. It is conservatively listed under English and the general multilingual overview until a named-language source is confirmed.
Official language coverage ↗Where it helps
Compact search indexes
Represent short text collections with smaller default vectors.
Interactive discovery
Evaluate a lighter model for frequently used search experiences.
Supported-language content
Retrieve help or product text in the languages covered by Nano.
When to choose it
Consider it when its language coverage fits your audience and you want a compact Jina model. Compare Nano with Small on representative questions, especially when documents or languages are varied.
Awaiting provider activation and verified account pricing. This site currently lists the same service rate for Nano and Small; a smaller original model is not a promise of a lower hosted price.
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