README.md

---
title: README
colorFrom: gray
colorTo: indigo
sdk: static
pinned: false
---

# Build on Apertus 🇨🇭

Welcome to **Hack Apertus**, an open-source hackathon movement built around **Apertus**, the fully open large language model. Hack Apertus is a three-stage hackathon series with an Online Hackathon and six On-site Hackathons in Swiss cities, running from October 2026 until May 2027. Interdisciplinary teams build on Apertus to create reusable blueprints deployable air-gapped, on-premise, or on sovereign cloud infrastructure. You can join any hack, or all of them. Thanks for joining and contributing your time and expertise. 👋

## What we build on: Apertus

Apertus is Switzerland's fully open LLM, developed by **EPFL, ETH Zurich and CSCS** within the **Swiss AI Initiative (SNAI)**. It is open weights + open data + full training recipes, trained on 1,800+ languages (incl. Swiss German and Romansh), and developed with the EU AI Act in mind. First released 2 September 2025 under Apache 2.0; the 1.5 versions (8B/70B) followed on 24 July 2026.

Hack Apertus builds on the **v1.5** family: continued pretraining of Apertus 1.0 that adds multimodal input, tool use, and up to 262k-token context.

## 🗓️ Hackathon Roadmap

| Stage | Dates | Details |
| :--- | :--- | :--- |
| **01. Online Hackathon** | Oct 1–16, 2026 | Open to everyone worldwide. Use case prototypes, model adoption, and academia challenges |
| **02. On-site Hackathons** | Oct 2026 – May 2027 | 2-day local hacks in **Bern, Zurich, Basel, Biel/Bienne, Lausanne & St.Gallen** |
| **03. Grand Finals** | May 14, 2027 | Closing showcase and pitch competition of all Hack winners for the main «Hack Apertus» Award in **St.Gallen** |


## Which models to use

| Model |
| :--- |
| [`swiss-ai/Apertus-v1.5-8B`](https://huggingface.co/swiss-ai/Apertus-v1.5-8B) |
| [`swiss-ai/Apertus-v1.5-70B`](https://huggingface.co/swiss-ai/Apertus-v1.5-70B) |
| [Apertus 1.5 multimodal](https://huggingface.co/swiss-ai) |
| [`swiss-ai/Apertus-8B-Instruct-2509`](https://huggingface.co/swiss-ai/Apertus-8B-Instruct-2509) |

> v1.5 weights are gated. Log in on Hugging Face and accept the conditions to download. Deployment is supported in Transformers and vLLM.

## Useful links

**Apertus & the Swiss AI Initiative**

- 🤗 [swiss-ai on Hugging Face](https://huggingface.co/swiss-ai) · all models, datasets, papers
- 📚 [Apertus LLM collection](https://huggingface.co/collections/swiss-ai/apertus-llm-68b699e65415c231ace3b059)
- 🌐 [apertus-ai.org](https://www.apertus-ai.org/) · official site, documentation, contact
- 💻 [github.com/swiss-ai](https://github.com/swiss-ai) · code, legal, website source
- 📄 [Apertus tech report (arXiv 2509.14233)](https://arxiv.org/abs/2509.14233)

**Run & access Apertus**

- ⚙️ [vLLM](https://github.com/vllm-project/vllm) · [Transformers](https://github.com/huggingface/transformers)

## Get involved

🌐 [hackapertus.ch](https://hackapertus.ch) · ✉️ [hello@hackapertus.ch](mailto:hello@hackapertus.ch) · 💼 [LinkedIn](https://www.linkedin.com/company/hack-apertus)

*Initiated by aiLights, Artifact and Swiss Engineering.*