EmbeddingGemma 2 is a 740-million-parameter open multimodal embedding model built on the Gemma 4 architecture and released under an Apache 2.0 license. It unifies text, code, images, video, and audio in a shared embedding space, and is designed for on-device inference, requiring as little as ~191MB active RAM for text-only weights on a Google Pixel 11 Pro. It improves MTEB Code scores from 68.76 to 78.68 over EmbeddingGemma and offers an 8K token context window.
No score is assigned. Sources and their independence are shown in the citation chain below.