Switzerland’s Apertus is a free, open-source rival to ChatGPT

Written by

Published 4 Sep 2025

Fact checked by

We maintain a strict editorial policy dedicated to factual accuracy, relevance, and impartiality. Our content is written and edited by top industry professionals with first-hand experience. The content undergoes thorough review by experienced editors to guarantee and adherence to the highest standards of reporting and publishing.

disclosure

apertus open source rival chatgpt scaled

Switzerland released its own artificial intelligence (AI) model on Tuesday, treating AI like public infrastructure instead of leaving the technology to private companies like OpenAI and Google.

The country launched Apertus, a large language model built by three public institutions that anyone can download and use for free. Swiss researchers designed the system to compete with ChatGPT and similar tools while keeping full control over the technology within their borders.

    “Currently, Apertus is the leading public AI model: a model built by public institutions, for the public interest,” said Joshua Tan, who works on making AI a public resource. “It is our best proof yet that AI can be a form of public infrastructure like highways, water, or electricity.”

    The Swiss Federal Institute of Technology Lausanne (EPFL), ETH Zurich, and the Swiss National Supercomputing Centre spent over 10 million computing hours creating the model. They trained it on 15 trillion pieces of text in more than 1,000 languages.

    Unlike most AI companies, Swiss developers made everything public. Users can see the computer code, training data, and exact methods used to build Apertus. Most tech companies keep these details secret to protect their business edge.

    “With this release, we aim to provide a blueprint for how a trustworthy, sovereign, and inclusive AI model can be developed,” said Martin Jaggi, a machine learning professor at EPFL.

    The model comes in two sizes. The smaller version has 8 billion parameters for individual users. The larger 70 billion parameter version compares to Meta’s Llama 3 from 2024.

    Swiss researchers focused heavily on languages beyond English. Forty percent of the training data uses non-English languages, including Swiss German and Romansh. Most AI models struggle with these smaller language communities.

    The project follows strict European rules about data privacy and copyright. Swiss developers only used publicly available information and respected website requests to block data collection. Many AI companies face lawsuits for using copyrighted material without permission.

    “Apertus is built for the public good,” said Imanol Schlag, the technical lead at ETH Zurich. “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.”

    Swiss banks and other industries welcome having local AI options. The Swiss Bankers Association sees “great long-term potential” for a model that follows Switzerland’s strict banking and privacy laws.

    However, Swiss companies already use other AI tools for their work. Adoption will depend on how well Apertus performs compared to established competitors like ChatGPT and Claude.

    The team plans regular updates and hopes to create specialized versions for law, healthcare, climate science, and education. They see this launch as the start of a long-term commitment to public AI development.

    Anyone can access Apertus through the AI platform Hugging Face or Switzerland’s Swisscom network.