---
title: "Google releases EmbeddingGemma 2, a 740M open multimodal embedding model"
url: https://www.parallelquant.com/posts/google-releases-embeddinggemma-2-a-740m-open-multimodal-embedding-model-b38b75
source_name: "The Decoder"
source_url: https://the-decoder.com/google-claims-embeddinggemma-2-outperforms-rival-embedding-models-twice-its-size/
published: 2026-10-06T19:47:27.000Z
topics: ["open source", "llms"]
publisher: "Parallel Quant"
---

# Google releases EmbeddingGemma 2, a 740M open multimodal embedding model

*2026-10-06 · Source: [The Decoder](https://the-decoder.com/google-claims-embeddinggemma-2-outperforms-rival-embedding-models-twice-its-size/)*

EmbeddingGemma 2 has 740 million parameters and converts text, images, video, audio and code into vectors. It needs about 191 MB of RAM, runs on-device, and Google says it outperforms some models twice its size. MarkTechPost reports it is built on Gemma 4 and ships under Apache 2.0.

**Why it matters:** A small, permissively licensed multimodal embedder makes fully offline retrieval-augmented generation (RAG) practical when paired with a small open model like Gemma 4. It fits the broader push toward private, on-device AI, and removes a common reason apps send data to external APIs.

**Topics:** open source, llms

---
Read the original: https://the-decoder.com/google-claims-embeddinggemma-2-outperforms-rival-embedding-models-twice-its-size/
Canonical: https://www.parallelquant.com/posts/google-releases-embeddinggemma-2-a-740m-open-multimodal-embedding-model-b38b75
Published by Parallel Quant — https://www.parallelquant.com