embedding-model

1 article
sort: new top best
clear filter
0 1/10

Google released Gemini Embedding 2, a natively multimodal embedding model that maps text, images, video, audio, and documents into a unified embedding space for RAG, semantic search, and classification tasks. The model supports flexible output dimensions via Matryoshka Representation Learning and integrates with multiple vector database and AI frameworks.

Gemini Embedding 2 Google DeepMind Gemini API Vertex AI LangChain LlamaIndex Haystack Weaviate QDrant ChromaDB Vector Search
blog.google · panarky · 3 days ago · details · hn