jina-embeddings-v5-text-nano

Name: jina-embeddings-v5-text-nano
Author: jinaai

by jinaai

Smallest high-quality multilingual text embedding at 239M parameters

176Kdl/month

97likes

212Mparams

HuggingFace Run it on your own data, free

Identifiers

Model ID

jinaai/jina-embeddings-v5-text-nano

Feature URI

mixpeek://text_extractor@v1/jina_embeddings_v5_nano_v1

Overview

Jina Embeddings v5 Text Nano is a 239M-parameter multilingual text embedding model built on the EuroBERT-210M backbone. It achieves 71.0 on MTEB English v2: remarkably close to the 677M v5-text-small (71.7) at one-third the size. Trained via embedding distillation from Qwen3-Embedding-4B with task-specific contrastive losses, it retains quality under aggressive dimension truncation and binary quantization.

On Mixpeek, jina-embeddings-v5-text-nano is the optimal choice for latency-critical and edge deployments where every millisecond counts. Its Matryoshka support (768 down to 32 dimensions) and robust quantization make it ideal for high-throughput text search at minimal compute cost.

Architecture

EuroBERT-210M backbone with last-token pooling. 239M parameters. Four task-specific LoRA adapters (retrieval, text-matching, clustering, classification). 8192-token context length. Matryoshka truncation from 768 to 32 dimensions.

Mixpeek SDK Integration

// No extractor parameter takes a Hugging Face model id (checked against
// GET /v1/discovery/extractors, which returns 13), so jina-embeddings-v5-text-nano runs
// on your side and the output is upserted through POST
// /v1/namespaces/{namespace_id}/documents/upsert. On Enterprise the other
// path is to upload the weights instead: POST /v1/namespaces/{id}/models
// accepts the huggingface format and a custom plugin loads them.
const res = await fetch(
  "https://api.mixpeek.com/v1/namespaces/ns_your_namespace/documents/upsert",
  {
    method: "POST",
    headers: {
      Authorization: "Bearer API_KEY",
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      collection_id: "col_your_collection",
      documents: [
        {
          document_id: "asset-00412",
          // The vector name has to match a vector index on the collection.
          vectors: { "text-embedding": yourVector },
          payload: { source_key: "archive/2026/asset-00412" },
        },
      ],
    }),
  },
);

// Managed alternative, if this exact model is not the requirement:
// text_extractor@v1 runs intfloat/multilingual-e5-large-instruct
// (1024-d) over a bucket, with no inference of your own.