Kolibri 1 BF16
Kolibri 1 BF16 is an open-weight mixture-of-experts reasoning model focused on German and English. It supports explicit reasoning, tool calling, and contexts up to 1,048,576 tokens, and is served using Aleph Alpha's inference package with vLLM.
What it does
Use it to build German- and English-language assistants and systems for reasoning, coding, document processing, retrieval-augmented generation, and tool-driven workflows.
- Availability
- Release date: 3rd of October 2026; available as Aleph-Alpha/Kolibri-1-BF16
- Pricing model
- Not established
- Deployment
- Run with Transformers or vLLM; model memory footprint is about 156 GB for BF16 weights
- Privacy
- Not established
- Licence
- Apache 2.0
Capability
Modality
Access
Deployment
Pricing
Target user
Evidence
6 claims- Kolibri 1 is a mixture-of-experts model with 78B total parameters and 3.46B active parameters per token. [1]
- The model supports German and English, explicit reasoning mode, and tool calling. [1]
The model supports an explicit reasoning mode
Hugging Face. Quoted text is outside CC BY 4.0 and keeps its original rights. - Its context length is 1,048,576 tokens, with serving and complex-task use recommended at up to 262,144 tokens. [1]
| Context length | `1,048,576 tokens`; we recommend ≤262,144 tokens for serving efficiency and complex tasks |
Hugging Face. Quoted text is outside CC BY 4.0 and keeps its original rights. - The model is released as Aleph-Alpha/Kolibri-1-BF16 and is supported natively by Transformers from version 5.20. [1]
- The model can be served with the Aleph Alpha vLLM plugin using the documented package or container. [1]
- The model is intended for human-reviewed use and not as an autonomous component that acts without review. [1]
Sources
1
1
Hugging FaceOwner-controlled
https://huggingface.co/Aleph-Alpha/Kolibri-1-BF16
Retrieved 2026-10-28 00:39 UTC
Verified
Versions
4| Version | Effective | Change | Changed fields | State |
|---|---|---|---|---|
| 4 | 2026-10-21 00:08 UTC | material change | access, pricing, availability, deployment | Current |
| 3 | 2026-10-17 02:02 UTC | material change | capability, pricing, availability, deployment | Current |
| 2 | 2026-10-04 13:57 UTC | material change | capability, deployment, availability, deployment | Current |
| 1 | 2026-10-03 14:50 UTC | publication | None | Current |
In briefs
2- Backburner leads 25 AI developments in week 40, 2026 Weekly brief 2026-W40
- Microsoft introduces FrogNano-4B-2609 for repository-level coding workflows Daily brief 2026-10-03
Original catalog content is licensed under CC BY 4.0, attributed to wowsuchai. Source quotations and linked content keep their original rights.