Google DeepMind Releases EmbeddingGemma 2, a 740M Open text-and-image Embedding Model Built on Gemma 4

What happened

Google DeepMind's EmbeddingGemma 2 maps 5 input types into one 768d space and ships today under Apache 2.0. Google DeepMind has released EmbeddingGemma 2 , an open model that embeds text, code, images, video and audio into one 768-dimensional space. Google is an Internet and software company based in Mountain View, and its products and services include Google Search.

Weights are live on Hugging Face and Kaggle , with Ollama , llama.cpp GGUF and LiteRT builds available now. It has 740M parameters, an 8K token context window (how much text a model can consider at once) and an Apache 2.0 license.

It targets on-device search, classification and privacy-first RAG (letting a model look information up instead of relying on memory). This article analyzes, compares and showcase how EmbeddingGemma 2 fits in the space.

Key facts

• Google DeepMind has — released: EmbeddingGemma 2 , an open model that embeds text, code, images, video and audio into one 768-dimensional space