StockDuty ← Dashboard Oct 07, 2026 10:55 PM ET
SignalTech/AI

EmbeddingGemma 2: an open, lightweight multimodal embedding model

New open lightweight multimodal embedding model signals continued AI model release activity.

What happened: Google launched EmbeddingGemma 2, an open multimodal embedding model built on the Gemma 4 architecture and released under an Apache 2.0 license. It unifies code, images, video, and audio in a shared embedding space. Key numbers: 740 million parameters. 20 million downloads of EmbeddingGemma 1. Why it matters: The model targets on-device inference, letting developers build cross-modal search and retrieval that works offline. Because it shares Gemma 4's text tokenizer and audio encoder, both models can run in a unified pipeline with a lower combined memory footprint.
Google launched EmbeddingGemma 2, an open multimodal embedding model built on the Gemma 4 architecture and released under an Apache 2.0 license. It unifies code, images, video, and audio in a shared embedding space.
The model targets on-device inference, letting developers build cross-modal search and retrieval that works offline. Because it shares Gemma 4's text tokenizer and audio encoder, both models can run in a unified pipeline with a lower combined memory footprint.
Sources