Aleph Alpha has released Kolibri, a German-English language model with 78 billion parameters. About three billion of those parameters are active per token through a mixture-of-experts architecture.
Kolibri balances performance and cost
Aleph Alpha claims Kolibri sits on the Pareto front of quality and operating cost in both languages. The model outperforms compared models with similar architectures, including some significantly older models from March and April 2026, on either metric.
German accounts for 21.3 percent of the training data. This is backed by a dedicated German data pipeline the company built for the project. Chinese models were also used to generate synthetic training data.
The model targets specific public and industrial sectors
Aleph Alpha says Kolibri targets public administration, aviation, and industry. The model was developed under European law with the EU AI Act in mind.
Kolibri supports context windows of up to one million tokens. The tech report states the system was trained on 768 B200 GPUs located in Germany and Finland.
The weights are available under an Apache 2.0 license on Hugging Face.
