embeddinggemma-300m

Name: embeddinggemma-300m
Author: google

by google

Google's efficient multilingual text embedding -- 100+ languages in 300M parameters

1.5Mdl/month

308Mparams

HuggingFace Run it on your own data, free

Identifiers

Model ID

google/embeddinggemma-300m

Feature URI

mixpeek://text_extractor@v1/google_embeddinggemma_300m_v1

Overview

EmbeddingGemma is Google's compact text embedding model based on the Gemma 3 architecture, designed for deployment on edge devices while maintaining strong multilingual performance. At 308M parameters, it runs in under 200MB of RAM when quantized and achieves sub-22ms inference on EdgeTPU.

Despite its small size, it ranks as the highest-performing open multilingual text embedding model under 500M parameters on MTEB. It supports Matryoshka representations (768 and 128 dimensions) for flexible memory/quality tradeoffs. On Mixpeek, it provides a lightweight embedding option for high-throughput text indexing where GPU resources are limited.

Architecture

Gemma 3 decoder backbone, 308M parameters. 2048 token context. Matryoshka dimensions: 768 (full) and 128 (compressed). Distilled from a larger teacher model. Optimized for TPU/mobile inference.

Mixpeek SDK Integration

// No extractor parameter takes a Hugging Face model id (checked against
// GET /v1/discovery/extractors, which returns 13), so embeddinggemma-300m 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.