Google DeepMindLaunchAI ModelsOctober 6, 2026

Google DeepMind releases EmbeddingGemma 2, a 740M multimodal embedding model

Read original source →blog.google

EmbeddingGemma 2 maps text, code, images, video, and audio into one 768-dimensional space under Apache 2.0, built on Gemma 4. It scales from 270M parameters for text-only to 740M with vision and audio encoders, using ~191MB active RAM on a Pixel 11 Pro.

7 sources

More stories today

Open the live feed