Skip to content
IBM open-sources a 32K context embedding model that beats everything under 100M parameters

IBM open-sources a 32K context embedding model that beats everything under 100M parameters

6 min read AI Data Architecture

IBM just released Granite Embedding Multilingual R2 under an Apache 2.0 license. With a 32K context window, it delivers state-of-the-art retrieval quality for models under 100M parameters, making highly efficient, local RAG applications easier and cheaper to build....

Subscribe to listen

/