Facts
- Announced
- EmbeddingGemma 2, an embedding model that maps text, code, images, video, and audio into one vector space · 2026-10-07
- Release
- game · 2026-10-07 · 2026-10-07
- Noted
- It has 740 million parameters, an 8K-token context window, and output vectors that developers can shorten from 768 dimensions to 512, 256, or 128. · 2026-10-07
- Noted
- Google says text-only weights need about 191MB of active RAM on a Pixel 11 Pro. · 2026-10-07
- Noted
- The pitch is local: a multimodal embedder small enough to hold its working memory to roughly 191MB for text on a phone is aimed at on-device search and retrieval-augmented generation, where the vector index sits on the hardware instead of a server. · 2026-10-07
Connections
Structured graph also available as JSON at /public/entities/embeddinggemma-2.
CC BY 4.0.
1h ago
Google announced EmbeddingGemma 2, an embedding model that maps text, code, images, video, and audio into one vector space. It has 740 million parameters, an 8K-token context window, and output vectors that developers can shorten from 768 dimensions to 512, 256, or 128. Google says text-only weights need about 191MB of active RAM on a Pixel 11 Pro.