Switzerland has unveiled Apertus, a multilingual Large Language Model (LLM) developed by ETH Zurich, EPFL, and the Swiss National Supercomputing Centre.
Unlike proprietary AI models, Apertus is fully open—its source code, training data, and model weights are available for anyone to explore on Hugging Face.
Transparency by Design: Apertus reveals how it was trained and on what data, making it one of the most auditable AI systems available.
Multilingual Scope: Trained on 15 trillion tokens across 1,000+ languages, including underrepresented ones like Romansh and Swiss German.
Ethical Compliance: Only public data was used, respecting copyright and EU AI regulations.
Accessible to All: Released in 8B and 70B parameter versions, ready for both researchers and industry players.
Why it matters: Apertus could reshape the AI landscape by offering an ethical, transparent, and sovereign alternative to closed corporate models. For regulators, businesses, and researchers, this represents a benchmark in trust-first AI development.
Source: The Verge