The AI model includes many languages that have so far been underrepresented in LLMs, such as Swiss German and Romansh. Switzerland has joined the AI race with the launch of its new large language model (LLM), Apertus, developed as part of the Swiss AI initiative and led by universities EPFL and ETH Zurich, as well as the Swiss National Supercomputing Centre. Trained on 15trn tokens across more than 1,000 languages, Apertus has utilised languages that, according to the developers, have been vastly underrepresented in the LLM space, such as Swiss German and Romansh. The team said that Apertus has been created with due consideration given to Swiss data protection laws and copyright laws and the transparency obligations under the EU AI Act. This comes at a time when AI developers are being heavily criticised for the methods they use to gather, deploy and store information for training their systems.