On the 9th of July 2025, ETH Zurich announced in a news article its new large language model, which is being developed in collaboration with the Swiss National Supercomputing Centre (CSCS). The release is planned for late summer.

The focus is on a consistent open-source approach: model architecture and training data should be openly accessible and transparently documented. In this way, ETH is setting an example for trustworthiness, reproducibility and social benefit.

The model was trained using the new Alps supercomputer, one of the most powerful in Europe. It will be released in two sizes – with 8 and 70 billion parameters. According to ETH Zurich, the larger version will rank among the most advanced open-access language models worldwide. The model supports over 1,000 languages and s additionally optimized for applications in programming and mathematics.

We welcome this open and responsible approach. As a technology-driven company, we are following these developments with great interest – and look forward to the release later this summer 2025.