Gemini Embedding 2 Now Generally Available in Gemini API and Vertex AI

Gemini Embedding 2 Now Generally Available in Gemini API and Vertex AI

Google has announced the general availability of Gemini Embedding 2 in the Gemini API and Vertex AI. Gemini Embedding 2 is the first natively multimodal embedding model from Google — capable of generating embeddings across text, image, and other modalities in a unified representation space. The GA release includes production-grade stability improvements and performance optimisations required for enterprise and production application deployments.

Why It Matters

Natively multimodal embeddings enable retrieval and semantic search across mixed-media corpora without the separate text-and-image pipeline workarounds that current RAG architectures require. For teams building on Google Cloud, GA availability means Gemini Embedding 2 can now underpin production search, recommendations, and retrieval systems — removing the barrier of preview instability that blocked enterprise adoption.