Universal embedding model for multimodal and multilingual retrieval. Supports both single-vector and multi-vector embeddings with task-specific LoRA adapters.
Modalities
Dimensions
2K
Max tokens
33K
Parameters
3.8B
Price / 1M tokens
$0.050
Type
🌍 Multilingual support (29+ languages)
| MTEB EN | 55.97 |
| MMTEB | 66.49 |
| COIR | 71.59 |
| LONGEMBED | 67.11 |
| VIDORE | 90.17 |
| JINA VDR | 84.11 |
| Release date | 2025-06-24 |
| License | Apache 2.0 |
| Model ID | jina-embeddings-v4 |
| Provider | Jina AI |
| Input image size | 768×28×28 |
Get detailed specifications for Jina Embeddings v4, including output dimensionality of 2048 dimensions, maximum input token length, supported input modalities, and pricing per million tokens. This embedding model is designed for semantic search, text classification, clustering, and retrieval-augmented generation applications where understanding the relationship between texts is essential.
Review the pricing structure for Jina Embeddings v4 and compare it against other embedding models from Jina AI and competitors. Understanding embedding model costs is essential when scaling vector search and RAG applications to millions of documents. We provide transparent pricing to help you budget effectively.
Explore the ideal use cases for Jina Embeddings v4. Whether you are building a semantic search engine, recommendation system, document classification pipeline, or multilingual retrieval system, understanding this model's capabilities and dimensionality will help you choose the right embedding strategy for your project.