Swiss National AI Initiative (SNAI) has released Apertus, Switzerland’s first large-scale open, multilingual language model — a milestone in generative AI for transparency and diversity. Trained on 15 trillion tokens across more than 1,000 languages – 40% of the data is non-English – Apertus includes many languages that have so far been underrepresented in LLMs, such as Swiss German, Romansh, and many others. Apertus serves as a building block for developers and organizations for future applications such as chatbots, translation systems, or educational tools.
The model is named Apertus – Latin for “open” – highlighting its distinctive feature: the entire development process, including its architecture, model weights, and training data and recipes, is openly accessible and fully documented.
This quickstart will help you get up & running with the Swiss LLM models built by the Swiss National AI Initiative (SNAI), on Azure.
Note: As the models are very new, this is a living documentation and will be updated frequently as more information gets publicly available about the models. If you find any errors or inconsistencies, please open a pull request.
We provide several ways to host the model on Azure:
| Azure Host Service | Model version | Status |
|---|---|---|
| Azure Virtual Machine with GPUs | 8b, 70b | |
| Azure Container Apps with Serverless GPUs | 8b | 🚧 Work in Progress |
| Azure Kubernetes Service with GPUs | 8b, 70b | 🚧 Work in Progress |
| Docker Image Apertus with vLLM | 8b, 70b |
Based on your requirements in terms of performance and control and on the preferred architecture, you can choose what works better for you.
@misc{swissai2025apertus,
title={{Apertus: Democratizing Open and Compliant LLMs for Global Language Environments}},
author={Apertus Team},
year={2025},
howpublished={\url{https://huggingface.co/swiss-ai/Apertus-70B-2509}}
}
 Francesco SODANO 📢 |
 Dominique BROEGLIN 📢 |