Cohere released North Small Translate on [date], saying it is the company's first translation model in the North model family. It is the company's first translation model update since the Command A Translate model.
Cohere reported an 83.6 score on WMT26 benchmarks, measured across 50+ languages. That compares with 81.56 for Qwen 3.5 397B A17B and 76.50 for GLM 5.2 FP8.
North Small Translate is built on a mixture-of-experts architecture and targets high-quality machine translation for global use. Availability begins today, initially for research and non-commercial use.
"North Small Translate marks a significant milestone as Cohere's first translation model in the North model family," said [name and title exactly as the source gives them]. The model represents a clear next step in Cohere’s commitment to offering high-quality machine translation wherever it is needed.
The announcement follows Cohere's expansion into multilingual and translation technologies. Cohere’s mission to make sovereign AI a technological reality is further advanced with this release.
Cohere did not say how the model will perform in real-time applications, and [the open question or limitation the source raises]. The model is now available for research and non-commercial use.
Source: cohere
