Switzerland releases its own AI model trained on public data

Switzerland has launched an open-source AI model called Apertus as an alternative to proprietary models like OpenAI's ChatGPT or Anthropic's Claude. The model's source code, training data, model weights, and detailed development process are available on the AI model platform HuggingFace. Apertus, which means "open" in Latin, was designed to set a new standard for trustworthy and globally relevant open models. The model was trained on over 1,800 languages and comes in two sizes with either 8 billion or 70 billion parameters, comparable to Meta's Llama 3 model. The model was built to adhere to the European Union's copyright laws and voluntary AI code of practice, which some US-based AI companies have been reluctant to sign, claiming that the regulations will curb AI innovation and deployment. Apertus' training data was restricted to public sources, and the developers adhered to the AI crawler opt-out requests on certain websites.