Swiss AI Initiative releases Apertus, a fully open foundation model built for EU sovereign AI
Tags AI · OSS · Infrastructure
The Swiss AI Initiative, a collaboration between EPFL, ETH Zurich, and CSCS, released Apertus, a fully open foundation model designed for "sovereign AI" deployment. The model comes in 8B and 70B parameter versions, is competitive with top open models at equivalent scale, and is multilingual from day one — trained on 1,000+ languages. Unlike most open-weight models, Apertus makes training data, code, weights, methods, and alignment principles fully documented and reproducible. It is built to comply with the EU AI Act, respecting opt-outs, removing PII, and preventing memorization. Swisscom is a strategic partner, offering Apertus on its enterprise platform. The release positions Apertus as a European alternative to US and Chinese foundation models for organizations that need regulatory compliance and data sovereignty.
Technical significance
Apertus addresses a gap in the market for foundation models that are both fully open and designed for regulatory compliance. If it achieves competitive benchmarks at scale, it could become the default choice for European enterprises and governments that cannot use US or Chinese models due to data sovereignty requirements. The 1,000+ language training is notable for multilingual deployment scenarios.