Swiss AI Initiative

A 4B parameter instruction-tuned language model created via pre-training distillation from the Apertus-8B teacher model. Uses a dense transformer decoder with grouped-query attention and xIELU activations, trained on 1.…

• 💬 Chatting
• 💻 Coding
• 📝 Writing
• 🌍 Translations New

A 4B parameter open-source base language model created via pre-training distillation from the Apertus-8B teacher model. It uses a dense transformer with grouped-query attention and xIELU activations, natively supports 1…

• 📚 Large Language Models
• 🔍 Research
• 🌐 Multilingual communication
• 🧠 Model training New

A 1.5B parameter instruction-tuned language model built on a dense transformer decoder with grouped-query attention and xIELU activation. Created via pre-training distillation from an 8B teacher model trained on 1.7T to…

• 💬 Chatting
• 🌐 Text translation
• 📝 Writing
• 📝 Text editing

A 1.5 billion parameter base language model trained via distillation from an 8B teacher model. It supports 1,811 languages, uses a dense transformer architecture with grouped-query attention and xIELU activation, and ha…

• 📚 Large Language Models
• 🌍 Translations
• 🔄 Language model optimization

A 0.5B parameter multilingual instruction-tuned language model built via pre-training distillation from an 8B teacher model. Supports 1811 natively integrated languages on a dense transformer with grouped-query attentio…

• 📝 Text generation
• 🌐 Text translation
• 🗂️ Multilingual summarization

An 8B parameter open-source instruction-tuned language model pretrained on 15T tokens. It natively supports 1,811 languages, uses fully open data and weights with disclosed training recipes, and is designed to respect d…

• 💬 Chatting
• 🌐 Text translation
• 💻 Coding
• 📝 Writing
• 🔍 Research

A fully open 8B parameter language model by Swiss AI Initiative, pretrained on 15T tokens. It natively supports over 1800 languages including all Swiss national languages, uses only compliant open training data, and ach…

• 📝 Text generation
• 🌍 Translations ‍
• 💻 Code generation

Apertus Apertus 70B Instruct is a fully open 70B-parameter decoder-only language model trained from scratch on 15T tokens using a staged curriculum of web, code, and math data. It natively supports 1811 languages and a…

• 📝 Text generation
• 💻 Coding
• 🤖 Agents
• 🌍 Translations
• 🔢 Math