Kolibri-1 needs about 35.2 GB of memory at Q3_K_S (a 33.9 GB file) with an 8K context window, so it fits on a 48 GB GPU. Longer context needs more: at 32K tokens the KV cache alone adds about 3.4 GB.
Kolibri‑1 Q3_K_S GGUF – 78.1B MoE reasoning model
Kolibri‑1 is a 78.1‑billion‑parameter mixture‑of‑experts model optimized for German and English. The Q3_K_S quantized GGUF version is 33.87 GB and supports a native context of 262,144 tokens. It is designed for long‑context inference and reasoning tasks.
Good at
- Long‑context reasoning
- Multi‑step problem solving
- Tool‑calling in German and English
Limits
- Requires experimental llama.cpp patch
- Not validated beyond 4,096‑token context in this port
- Full model must reside in memory
Best for
- Document summarization
- Complex question answering
- Agentic workflow integration
Who it’s for. Developers building conversational assistants or internal knowledge systems that need German and English reasoning. The model is suitable for environments where output is reviewed before action.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q3_K_S | 33.9 GB | 35.2 GB | Too big | Get file → |
For 8 GB VRAM, the Q3_K_S quantization uses 35.2 GB at 8 k context, which exceeds the GPU memory. A larger GPU or CPU inference is required. If memory is limited, consider a higher‑bit quantization not listed here.
What GPU do you need for Kolibri-1?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: no quant fits at an 8K context.
- 6 GB VRAM: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: no quant fits at an 8K context.
- 24 GB VRAM: no quant fits at an 8K context.
- 32 GB VRAM: no quant fits at an 8K context.
- 48 GB VRAM: the best quant that fits is Q3_K_S (33.9 GB file, about 35.2 GB in use).
Which Kolibri-1 quant should you choose?
Q3_K_S (33.9 GB) is the middle option in this repository. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Kolibri-1
How much VRAM does Kolibri-1 need?
About 35.2 GB at Q3_K_S with an 8K context window. That is the 33.9 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run Kolibri-1 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 35.2 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Kolibri-1 GGUF?
It depends on the quant: Q3_K_S is 33.9 GB.
What context length does Kolibri-1 support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.8 GB at 8K and 3.4 GB at 32K.
Where can I download Kolibri-1 GGUF?
The files are on Hugging Face in the Eliasfpv28/Kolibri-1-Q3_K_S-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does Kolibri-1 use?
Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run Kolibri‑1 Q3_K_S on 8 GB VRAM?
No. The 33.87 GB GGUF file requires more than 8 GB of VRAM even at 8 k context.
Is the model fully supported by llama.cpp?
It requires the included Kolibri1 llama.cpp source patch; the base llama.cpp revision does not support this architecture.
What context length can I use?
The GGUF port has been tested locally only at 4,096 tokens; longer contexts need additional KV‑cache memory and validation.