← back to the directory
Model Retrieval & Memory

Jina Embeddings v3

1Multilingual 2embeddings 3with 4task-specific 5LoRA 6adapters

the six Ws · specification

W1 Who

Jina AI released Jina Embeddings v3.

W2 What

Jina Embeddings v3 is a multilingual, long-context text embedding model that uses task-specific LoRA adapters to optimize for retrieval, classification, or clustering.

W3 Where

Hosted on Hugging Face under jinaai/jina-embeddings-v3, with an accompanying API on the Jina AI website.

W4 When

Released in September 2024.

W5 Why

It supports up to 8192 token contexts across nearly 100 languages, targeting production retrieval-augmented generation pipelines.

W6 With

Runs via the Hugging Face transformers library with custom code enabled.

W7 Watch

Released under the CC-BY-NC 4.0 license for the open weights, restricting commercial use without a paid Jina AI license.

embeddingsmultilinguallong-contextJina AI

for agents & scripts

Reading this as a machine? Query it directly.

Search is open JSON - no key. Report telemetry after using a tool and it feeds that tool’s Proof Score. Or speak MCP to /mcp and discover tools mid-loop.