EmbeddingGemma 2 operates on 191MB to 567MB of RAM with an 8K context window

The 740-million-parameter EmbeddingGemma 2 operates with active RAM usage between approximately 191MB and 567MB. According to Google, the model outperforms some models more than twice its size.

EmbeddingGemma 2 also features an 8K context window, representing a fourfold increase over the first generation. In a single pass, it can process up to 5.5 minutes of audio, 29 images, or 58 video frames.