Granite Embedding Multilingual R2: Open Apache 2.0 Multilingual Embeddings with 32K Context — Best Sub-100M Retrieval Quality

IBM has released two new open-source multilingual embedding models, Granite Embedding Multilingual R2, under the Apache 2.0 license. The release includes a compact 97M-parameter model and a full-size 311M-parameter model, both built on ModernBERT. These models support over 200 languages, with enhanced retrieval quality for 52 languages and 9 programming languages. They feature a significantly expanded 32,768-token context length and are designed for easy integration into frameworks like LangChain and LlamaIndex for tasks such as retrieval-augmented generation.