Nemotron-3-Embed-1B-BF16

Name: Nemotron-3-Embed-1B-BF16
Author: nvidia

by nvidia

The 1B sibling of Nemotron-3-Embed, for when 8B does not fit the budget

1.14Bparams

HuggingFace Run it on your own data, free

Identifiers

Model ID

nvidia/Nemotron-3-Embed-1B-BF16

Feature URI

Overview

Same family and same retrieval-first intent as the 8B, at roughly an eighth of the size. The tradeoff is the usual one: recall on hard queries against throughput and cost per million chunks.

The useful way to choose is to run both over your own corpus rather than a benchmark. A 1B model that answers your queries is worth more than an 8B that answers someone else's, and the gap between them narrows sharply on domains where the vocabulary is already unambiguous.

Packaged for sentence-transformers and vLLM, like its larger sibling.

Architecture

Ministral3Model, 1,140,918,272 parameters, BF16 weights. Sentence-similarity pipeline packaged for sentence-transformers.

Mixpeek SDK Integration

// No extractor parameter takes a Hugging Face model id (checked against
// GET /v1/discovery/extractors, which returns 13), so Nemotron-3-Embed-1B-BF16 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.