https://derekwang2002.github.io/ai-intelligence/en/projects/kolibri/ Material changes in AI models, agents, engineering and infrastructure, with primary sources and recommendations. en Mon, 02 Nov 2026 16:56:00 GMT https://derekwang2002.github.io/ai-intelligence/en/events/ev-20261003-01/ ev-20261003-01-published Sat, 03 Oct 2026 16:01:57 GMT New TRIAL <p>Aleph Alpha released Kolibri under Apache-2.0, a German-English reasoning MoE with 78B total and 3.46B active parameters. It supports reasoning levels, tool calling and up to 1M tokens, while recommending at most 262K for efficient serving and complex tasks. Weights, a vLLM plugin and a detailed training report are public; evaluations are author-reported.</p><p><strong>Why read: </strong>Kolibri provides inspectable open weights for German enterprise RAG and tool calling.</p><p>Recommendation: TRIAL</p><p>Sources: <a href="https://huggingface.co/Aleph-Alpha/Kolibri-1">Kolibri-1 model card and weights</a> · <a href="https://aleph-alpha.com/en/blog/kolibri-has-landed-a-sovereign-open-weight-model/">Kolibri has landed</a> · <a href="https://github.com/Aleph-Alpha/aleph-alpha-inference">Aleph Alpha inference plugin</a></p>