Google DeepMind released EmbeddingGemma 2, an open embedding model for text and code, images, video, and audio. Its modular architecture has about 740M parameters and supports a shared 8,192-token input budget.
AI applications
EmbeddingGemma
For Developers building local retrieval, RAG and classification
Update history
1The releases collected here, ordered by announcement date. This is a selected history, not a complete changelog.
- vEmbeddingGemma 2
What it does
Google DeepMind’s lightweight embedding family. EmbeddingGemma 2 maps text, code, images, video and audio into a shared vector space, with downloadable weights and modular encoders for local deployment.
Who it’s for
Developers building local retrieval, RAG and classification
Materials & reading
Official introductions, docs, demos and press kits. These describe the product unless a specific release is shown.
- EmbeddingGemma 2: an open, lightweight multimodal embedding model
Original page reviewed
Open originalSource checked - EmbeddingGemma 2 model card
Original page reviewed
Open originalSource checked - google/embeddinggemma-2
Original page reviewed
Open originalSource checked
Coverage
This timeline contains selected, reviewed releases. It is not a complete history.
No source-check record is available for this product; recent coverage is unknown.
Report a correction
Tell us what needs correcting. Reports are reviewed before any public content changes.