Switzerland Launches ‘Apertus’, Its National Open-Source LLM

IBL News | New York

Switzerland entered the AI race with the launch of a national Large Language Model (LLM) named ‘Apertus’, intending to gain sovereign AI infrastructure as an alternative to American and Chinese dominance.

The model was developed by specialized engineers and a large number of researchers of ETH Zurich, the Swiss Federal Institute of Technology Lausanne (EPFL), and the Swiss National Supercomputing Centre (CSCS).

The model’s name, ‘Apertus,’ derived from the Latin word for “open,” highlights its open-source architecture, training data, model weights, and intermediate checkpoints. It comes in two sizes: 8 billion, for individual use, and 70 billion parameters.

The model, unveiled this week, is now accessible and documented on Hugging Face, serving as a building block for developers and organizations to create applications such as chatbots, translation systems, or educational tools.

As of today, only Swisscom business customers are able to access the Apertus model via its AI platform.

“We see it as a driver of innovation and a means of strengthening AI expertise across research, society, and industry,” said Thomas Schulthess, Director of CSCS and Professor at ETH Zurich.

“Apertus is built for the public good. It stands among the few fully open LLMs at this scale and is the first of its kind to embody multilingualism, transparency, and compliance as foundational design principles”, added Imanol Schlag, technical lead of the LLM project and Research Scientist at ETH Zurich.

The authors of this LLM said, without providing further detail, that “upcoming Swiss {ai} Weeks hackathons will be the first opportunity for developers to experiment hands-on with Apertus, test its capabilities, and provide feedback for improvements to future versions.”