Switzerland has officially unveiled its new national AI language model known as Apertus LLM. Developed by leading universities in the country, it aims to provide a transparent and inclusive alternative to models like ChatGPT and Llama.

Apertus LLM is compared in power to Meta's Llama 3 (2024), but Swiss developers emphasize that they do not seek to compete with major American or Chinese companies that invest billions in new developments. Instead, the focus is on code openness and security — allowing users to download the model to their own servers and maintain control over their data.

The model has already been dubbed a “new phase” in the history of open LLMs. Unlike many competitors, Apertus shares all technical details, including the “recipe” for training, which is expected to enhance trust and reduce risks of bias or “hallucinations.”

Apertus is anticipated to find applications in scientific research, healthcare, education, and even finance, where confidentiality and compliance with European standards are particularly crucial. The Swiss Bankers Association believes that this homegrown LLM has “significant long-term potential,” even though banks like UBS are already collaborating with OpenAI and Microsoft.

Despite some skepticism from the business community, the project has garnered support from the government and the scientific community as a strategic move to preserve digital sovereignty.