Overview of Soofi S 30B-A3B Model
We introduce 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, maintaining a near-constant inference cache as context grows, providing a decisive throughput advantage over dense models for long-context, high-concurrency deployment.
Pretrained on approximately 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, outperforming every European sovereign baseline in our comparison, including those with far larger active parameters.
Among fully open models, Soofi S obtains the highest evaluation scores in English and German, 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 will all be made public. Where source licenses permit, data-construction artifacts will be released under permissive licenses; commercially licensed sources are documented with aggregate statistics and exact mixture accounting.
Blogger's Review: The release of Soofi S 30B-A3B marks a significant advancement in the open-source foundation model space, especially showcasing the advantages in high concurrency and long context processing. Its impressive performance in both German and English underscores the effectiveness of language-specific optimizations, potentially paving the way for the development of foundational models for more countries and languages.