Gemini Embedding 2
Multimodal embeddings for advanced search and analysis
Gemini Embedding 2 is the first natively multimodal embedding model from Google, mapping text, images, audio, video and documents into a single embedding space. This enables more accurate retrieval and classification across different media types, unlocking powerful new use cases. Designed for efficiency and broad compatibility, it helps developers build advanced multimodal search and data analysis systems.