Kolibri 1
A solid pick for self-hosting and full data control.
Overview
Kolibri 1 was released on 3 October 2026. It is a mixture-of-experts model with 78.1 billion parameters, 3.46 billion of which are active per token, so its compute cost is close to that of a small model. Full weights are published on Hugging Face in BF16 and FP8 under the Apache 2.0 licence, which allows commercial use, modification and redistribution. Aleph Alpha targets public administrations and regulated industries that want to host it on their own infrastructure. The trained context is 262,144 tokens, with a configuration validated up to one million. Note for non-German users: the model is declared for German and English only, French is not part of its declared corpus. Published scores (GPQA Diamond 84.8%, LiveCodeBench v6 85.1%) are vendor-measured and not yet independently confirmed.
Skill profile
Not disclosed
Three independent measures
What people prefer, what the model can do, what the market uses. Each source has its biases and none is enough on its own: we show them side by side, never blended.
Strengths
- Open-source and self-hostable
Limitations
- No documented EU residency
- API pricing not disclosed
Who is it for
- you want to control cost or self-host
- you handle sensitive data with no EU option
Access & availability
Key specifications
What it costs per month
Estimate from sourced API prices and a hardware tier for self-hosting. Assumptions are editable.
Assumptions: 36-month amortisation, 0.25 €/kWh, 22 days/month, 1 $ = 0.92 €. Excludes labour, software licences, redundancy.
nAIvigate Trust Index
What the vendor offers and documents for professional use in Europe. An undocumented criterion counts as zero and stays visible: the score rises when the vendor publishes, not when we assume.
13/19 criteria documented · 68 % · 7 of them set by rule
See the criteria▾
Trust Index badge
Show this score in your docs, README or security page. It updates automatically.
[](https://naivigate.com/en/model/kolibri)Data & compliance
What we know, line by line, with source and date. These lines feed the Data and Enterprise families of the Trust Index above. GDPR compliance depends on your processing, not on the model alone.
| EU data residency | Not documented |
| Data processing agreement (DPA) | Not documented |
| Transfer outside the EU | Not documented |
| Training on your data | Not documented |
| Configurable retention | Not documented |
| Zero Data Retention | Not documented |
| Self-hostable | Yes |
| Declared hosting | self |
Quality of this card’s data
This score measures the quality of the data on this card: price, context, Elo, licence, documented, sourced and fresh. It measures neither the model nor its compliance: for that, see the Trust Index above.
4/8 documented · 2/8 primary sources
Run it locally
| Quantization | Disk | RAM / VRAM | Typical hardware |
|---|---|---|---|
| Q4 · recommended | 44.5 GB | 51 GB | 64GB Mac / dual 24GB GPUs |
| Q8 · balanced | 83.6 GB | 94 GB | 96-128GB Mac Studio / 4× GPUs |
| FP16 · max quality | 156.2 GB | 174 GB | Server GPU infra (H100/A100…) |
Estimates for a moderate context. Long contexts need more RAM (KV cache).
Deploy
Copy-ready commands generated from this card. Adjust context length and GPU count to your hardware.
OpenAI-compatible server for production on NVIDIA GPUs.
pip install vllm
vllm serve Aleph-Alpha/Kolibri-1 \
--max-model-len 32768 \
--tensor-parallel-size 2 \
--dtype autoAdvanced data · for expertsArchitecture, modalities, detailed cost, full benchmarks▾
| Arena Elo | — |
| MMLU | — |
| GPQA | — |
| HumanEval | — |
| SWE-Bench | — |
| MATH | — |