Cohere Releases North Small Translate: A 218B MoE Translation Model That Scores 83.6 on WMT26 Across 50 Languages

Cohere launched North Small Translate, a 218-billion-parameter Mixture-of-Experts model for translating 50 languages, achieving a score of 83.6 on the WMT26 benchmark. The model uses 25 billion parameters per token and is freely available for non-commercial use, with commercial access via Cohere Model Vault or RWS Language Weaver.