A Sovereign, Open-Source Foundation Model for German and English

2026-07-10

AI Summary

This work addresses the current lack of high-performance, open-source, and sovereignty-compliant foundation models tailored for German and English, particularly under long-context and high-concurrency conditions. We present Soofi S 30B-A3B, the first open-source sovereign mixture-of-experts (MoE) model designed for European languages, which integrates Mamba and Transformer architectures and activates only 3 billion of its 30 billion parameters during inference. Trained end-to-end on Germany’s industrial AI cloud using a 27 trillion token multilingual corpus with German weighting, the model matches or exceeds the performance of dense 14–27B models on comprehensive English-German benchmarks. It achieves state-of-the-art code generation capabilities among 17 open-source base models and sets a new record for bilingual evaluation scores among fully open models, significantly outperforming existing European sovereign alternatives.

Abstract

We present Soofi S 30B-A3B, a sovereign, open-source Mixture-of-Experts (MoE) hybrid Mamba Transformer foundation model for German and English. Its hybrid design activates only 3B of 30B parameters per token and keeps the inference cache near-constant as context grows, giving it a decisive throughput advantage over dense models for long-context, high-concurrency deployment. Pretrained on roughly 27 trillion tokens with deliberately up-weighted German, Soofi S matches dense 14 to 27B models on aggregate English and German benchmarks while achieving the best code aggregates in both languages among 17 open base models, and outperforms every European sovereign baseline in our comparison, including ones far larger in active parameters. Among fully open models, Soofi S obtains the highest English and German evaluation scores, ahead of Olmo 3 32B and Apertus 70B. Soofi S was built end-to-end on the German Industrial AI Cloud, a sovereign HPC scale AI infrastructure operated by Deutsche Telekom in Munich. Soofi S will be released under highly permissive, open-access terms: weights, selected intermediate checkpoints, full per-source data accounting, hyperparameters, and training and evaluation code. Where source licenses permit, data-construction artifacts are released under permissive licenses; commercially licensed sources are documented with aggregate statistics and exact mixture accounting.

Problem

• sovereign AI
• foundation model
• open-source
• German language
• multilingual modeling

Innovation

• Mixture-of-Experts
• Mamba-Transformer hybrid
• sovereign AI
• long-context inference
• open-source foundation model