Jina Embeddings v2 Base English
Jina Embeddings v2 Base English is a JinaBERT encoder with an 8,192-token input window. It creates pooled vectors for semantic search and similarity, rather than generating an answer.
How the model works
The backbone was pretrained on C4, then trained with more than 400 million sentence pairs and hard negatives. The embedding training used 512-token sequences; symmetric bidirectional ALiBi enabled longer input at inference.Jina Embeddings v2 Base English model details
Use and access
The Apache 2.0 weights use mean pooling over token representations. Its card documents local inference and historical hosted access. The v2 model remains available as a download even though later multilingual families exist. The separate Small English release reduces the backbone to 33 million parameters.Jina Embeddings v2 Base English model detailsJina Embeddings v2 Base English technical overview
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | |
| Access | Downloadable |
| Model identifiers | jina-embeddings-v2-base-enjinaai/jina-embeddings-v2-base-en |
| Context window | 8,192 tokens Maximum input sequence in the model card; serving limits can be lower. Jina Embeddings v2 Base English model details |
| Parameters | 137 millionJina Embeddings v2 Base English model details |
| Weight license | Apache 2.0Jina Embeddings v2 Base English model details |
| Base model | JinaBERTJina Embeddings v2 Base English model details |
Related models
- Jina Embeddings v3
Jina AI · Jina Embeddings v3
A multilingual text model with separate query and passage adapters.
Launch date API · Downloadable - Jina Embeddings v2 Small English
Jina AI · Jina Embeddings v2
A 33-million-parameter English encoder with long-document input.
Launch date Downloadable