IBM released Granite Embedding Multilingual R2, two Apache 2.0 embedding models built on ModernBERT, covering 200+ languages with 32K-token context. The 97M-parameter compact model scores 60.3 on MTEB Multilingual Retrieval, the highest among open sub-100M multilingual embedders, while the 311M model scores 65.2. Radu Florian, Parul Awasthy, Aashka Trivedi, and Madison Lee contributed to the release, which supports code retrieval across nine programming languages.
No score is assigned. Sources and their independence are shown in the citation chain below.