Skip to content

Jina Embeddings V2 Base

An English embedding model for long text

Visit official site
Category AI models

What Jina Embeddings V2 Base is

The model is an English text embedding model built on JinaBert that handles long document sequences up to 8192 tokens.

What you can do with it

  • Compute semantic text similarity
  • Retrieve and rerank long documents
  • Build a recommendation system
  • Support RAG and generative search

Model notes

  • Pretraining used the C4 dataset
  • Over 400 million pairs were used later
  • 137 million parameters are reported
  • Single GPU inference is recommended

Pros & cons

✓ What we like

  • 8192 sequence length is supported
  • RAG use cases are named

! What to watch out for

  • It targets English text
  • Single GPU inference is recommended

FAQ

How long is the context?

The material says 8192 tokens.

How large is it?

The material says 137 million parameters.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review