# Colibri v2.1 adds native GGUF support

> Native GGUF support spans three inference backends.

By BIGCHANGE.AI Editorial

Published: 2026-11-02T15:48:38.773Z
Updated: 2026-11-02T15:48:38.773Z
Canonical: https://bigchange.ai/blog/colibri-v2-1-native-gguf

![Conceptual charcoal illustration of an unbranded compact desktop computer with its cover removed, exposing a circuit board and a small orange connection between memory packages and a compute chip.](https://bigchange.ai/api/media/file/local-model-shared-memory-workstation-hero-v1.png)
AI-generated conceptual editorial illustration by BIG CHANGE.

Colibri v2.1.0 adds native GGUF support across CPU, Vulkan and CUDA engines, the [project reported October 31](https://github.com/JustVugg/colibri/releases/tag/v2.1.0).

## The big change

- The maintainers say compatible published GGUF weights run without another conversion or quantization, removing a preparation step when switching to Colibri.
- A mixture-of-experts model selects only some components for an input, so its total parameter count does not tell the whole story about the work done on each request.
- Storage and memory tradeoffs shape which models are practical on a given machine.

## Why the format matters

[Hugging Face’s GGUF documentation](https://huggingface.co/docs/hub/gguf) describes it as a model format containing tensors and standardized metadata. It supports several numerical representations, including quantized types. Quantization changes how weights are represented, so a published file may already reflect a particular size and precision choice.

Colibri imports compatible files while retaining their existing blocks.

## Why streaming has limits

A [sparsely gated mixture-of-experts paper](https://arxiv.org/abs/1701.06538) describes a network that selects a subset of expert components for each input.

Colibri already read experts from storage as needed rather than keeping them all in memory, according to its documentation; streaming and single-GPU operation predate this release. The tradeoff is that the full model must still live on storage, and retrieving parts as needed makes storage access relevant to responsiveness.

The project’s [v2.1 documentation](https://github.com/JustVugg/colibri/blob/v2.1.0/README.md) lists 8 GB RAM minimum and 22 GB free disk for its smallest model; larger models need different resources. BIG CHANGE has not independently tested it.

## Sources & further reading

- [Colibri v2.1.0 release and documentation](https://github.com/JustVugg/colibri/releases/tag/v2.1.0) — Features and setup.
- [Hugging Face on GGUF](https://huggingface.co/docs/hub/gguf) — Format background.
- [Sparse mixture-of-experts paper](https://arxiv.org/abs/1701.06538) — Expert selection background.

## Sources

- [Colibri v2.1.0 release and documentation](https://github.com/JustVugg/colibri/releases/tag/v2.1.0) — Features and setup.
- [Hugging Face on GGUF](https://huggingface.co/docs/hub/gguf) — Format background.
- [Sparse mixture-of-experts paper](https://arxiv.org/abs/1701.06538) — Expert selection background.