Apertus 1.5: Open AI Model Understands Images & Audio

EPFL, ETH Zurich, and the Swiss National Supercomputing Centre have released Apertus 1.5. This is the latest version of their fully open large language model. What's interesting is Apertus 1.5 can now understand images and audio, in addition to text. It also offers improved reasoning and stronger instruction following. The model was trained on the Alps supercomputer and is one of the world's largest fully open language models. This means its training data, code, and development process are transparent. The long-term vision is to build a sovereign AI infrastructure, offering an open alternative to commercial models. Apertus is already being used in various applications, including an AI translation service for government documents in the Canton of Ticino. Researchers at EPFL have also used it for a medical large language model framework. In journalism, a Basel-based online news outlet uses Apertus to prepare local political news and for a search and analysis tool. This helps journalists explore parliament transcripts and trace politicians' positions. The bottom line is Apertus aims to provide trustworthy, transparent AI that organizations can build upon with confidence.