Swiss Institutions Unveil Fully Open AI Model, Apertus – Empowering Public with Multilingual, Transparent, and Sovereign AI
A groundbreaking open-source AI model, named Apertus, has been introduced by a consortium of Swiss institutions, including EPFL, ETH Zurich, and the Swiss National Supercomputing Centre (CSCS). The new model is designed to serve as a foundation for future research and applications, with its core principle being openness – every aspect of its design and training process is accessible to the public.
The versatile Apertus can be utilized by developers and organizations for creating chatbots, translation tools, education-focused applications, among other uses. It can be directly downloaded from Hugging Face or accessed through Swisscom, a strategic partner of the initiative. Two versions are available – an 8-billion-parameter model and a larger 70-billion-parameter version, both released under a permissive open-source license suitable for research, education, and commercial projects.
Apertus stands out among other AI systems by being fully transparent, with its architecture, training data, and documentation open to scrutiny. “With this release, we aim to provide a blueprint for how a trustworthy, sovereign, and inclusive AI model can be developed,” stated Martin Jaggi, Professor of Machine Learning at EPFL and member of the Steering Committee of the Swiss AI Initiative.
The team responsible for Apertus – comprising engineers and researchers from CSCS, ETH Zurich, and EPFL – will regularly update the model to ensure its continued relevance and effectiveness. Thomas Schulthess, Director of CSCS and Professor at ETH Zurich, described Apertus as “a driver of innovation and a means of strengthening AI expertise in research, society, and industry.”
The training process involved 15 trillion tokens in over 1,000 languages, with approximately 40% of the data coming from non-English languages. The model includes underrepresented languages like Swiss German and Romansh. Imanol Schlag, technical lead of the project and Research Scientist at ETH Zurich, emphasized that Apertus is “one of the few fully open LLMs (Language Models) at this scale and the first to embody multilingualism, transparency, and compliance as foundational design principles.”
Swisscom has already integrated Apertus into its sovereign AI platform, underscoring its commitment to fostering a secure and responsible AI ecosystem that serves the public interest and strengthens Switzerland’s digital sovereignty. Developers can test Apertus during the Swiss {ai} Weeks, which continue until October 5, 2025. International users will be able to access Apertus through the Public AI Inference Utility.
Apertus marks a significant step towards demonstrating that generative AI can be both powerful and open. “The release of Apertus is not a final step but rather the beginning of a journey, a long-term commitment to open, trustworthy, and sovereign AI foundations for the public good worldwide,” said Antoine Bosselut, Professor at EPFL and Co-Lead of the Swiss AI Initiative. Future updates will focus on expanding the model family, improving efficiency, and developing domain-specific tools for areas such as law, health, climate, and education while maintaining high standards of transparency.
