Google DeepMind Releases EmbeddingGemma 2, a 740M Open Multimodal Embedding Model Built on Gemma 4

Chronological Source Flow
Back

AI Fusion Summary

Google DeepMind has released EmbeddingGemma 2, an open 740M parameter multimodal embedding model built on Gemma 4. Licensed under Apache 2.0, this model maps text, code, images, audio, and video into a single 768d space. It is designed to run entirely on phones or small boards. While supporting 100 languages, Google notes that performance varies across them. The model has received no safety tuning and is available for immediate use by the developer community.
Community Comments
Loading updates...
0