Google DeepMind Launches EmbeddingGemma 2 for On-Device Multimodal Search
Google DeepMind's EmbeddingGemma 2 brings multimodal search to devices without cloud dependency — less memory, same capability.
Why it matters
A smaller, on-device embedding model for text, image, video and audio search reduces cloud reliance and memory footprint for consumer and edge applications. Practitioners building search or RAG systems can now evaluate local alternatives to cloud APIs.
The key facts
10 to knowEmbeddingGemma 2 supports text, image, video, and audio embeddings
Designed for on-device deployment with lower memory requirements
Reduces reliance on cloud AI infrastructure
Google DeepMind release (October 7, 2026)
Story key exists in recent catalog: google-embeddingemma-2-open-weight-release
EmbeddingGemma 2 supports text, image, video, and audio search
Designed for lower memory footprint and reduced cloud reliance
Open-weight model (inferred from 'Google DeepMind launches')
On-device execution capability
Story key already exists in last 72h: google-embeddingemma-2-open-weight-release
The story so far
Earlier coverage of this storyline
- Google DeepMind Releases EmbeddingGemma 2, a 740M Open Multimodal Embedding Model Built on Gemma 4MarkTechPost
- EmbeddingGemma 2: an open, lightweight multimodal embedding modelGoogle DeepMind Blog
- EmbeddingGemma 2Simon Willison
- Google expands EmbeddingGemma beyond text to images, audio and videoSiliconAngle
- This story
Go to the source
TechRepublictechrepublic.com
Publisher excerpt: Google DeepMind’s EmbeddingGemma 2 brings text, image, video and audio search to consumer devices with lower memory and less reliance on cloud AI.