Aleph Alpha Kolibri: How the sovereign German LLM works
- ID
- 31507
- Status
- summarized
- Published
- 03 Oct 2026, 6:43 PM
- Fetched
- 04 Oct 2026, 6:34 AM
- Provider
- Hacker News
- Category
- dev-community
- Original URL
- https://tej.as/blog/aleph-alpha-kolibri
- Source URL
- https://hnrss.org/best
Summary
- Score
- 7.0
- Created
- 04 Oct 2026, 6:34 AM
- Tags
- Audience
- developersai_ml_learnersai_agent_userssaas_founders
What happened
Aleph Alpha released Kolibri on 3 October 2026, an open-weight German/English mixture-of-experts LLM with 78.1B total parameters but only 3.46B active per token, under Apache 2.0 for the weights and config files (training code and methods stay proprietary). It was trained from scratch on ~24 trillion tokens — over a fifth German — on 768 NVIDIA B200 GPUs using infrastructure in Germany and Finland, with a 262,144-token native context (tested to 1,048,576), four reasoning levels, tool calling, a 18 June 2026 knowledge cutoff, and about 78 GB of FP8 weights. Aleph Alpha frames it as 'sovereign': built under European/German law with no foreign control, so customers get full deployment freedom and 'compliance as an inherited property', and it has signed the EU's GPAI Code of Practice. The 409-point Hacker News thread drew only 11 comments.
Why it matters
The ~78 GB FP8 footprint means Kolibri can plausibly run on a single 80 GB accelerator rather than a cluster, which is the concrete difference between self-hosting and paying per-token to a US API. If you sell into the EU, handle data that cannot leave a client's building, or need tool-calling agents with a 262k context window, this is a deployable alternative — but the 'scores above every compared model of its size in both languages' claim comes from Aleph Alpha's own evaluation, so benchmark it yourself before committing. For Malaysian and SEA builders, the relevant lesson is the packaging: weights + license + no-foreign-control deployment story as a compliance argument, which is a template local sovereign-model efforts can copy.
Discussion angle
Does 'sovereign' actually mean anything operationally? Test it against the details: Apache 2.0 covers weights but not training code, training ran on NVIDIA B200s, and the headline eval numbers are self-reported — ask what a builder would need to verify before swapping this in for a hosted API on a data-residency-sensitive workload.