---
title: "Model reference"
description: "The three Neutrino repositories, their sizes and measurements, and what each file is for."
canonical: "https://www.fermionresearch.com/docs/model-reference/"
source: "Fermion Research"
---

Docs / Reference

# Model reference

The three Neutrino repositories, their sizes and measurements, and what each file is for.

## Neutrino-1 8B

The default general model and the only Neutrino-1 model that reliably emits tool calls. The trained context window is 40,960 positions.

| Artifact | Size | Used by |
| --- | --- | --- |
| neutrino-8b_v4.bin | 3.88 GB | fermion native and PyTorch paths, plus MLX |
| neutrino-8b_v4.tv4z | 2.56 GB | Apple Silicon download transport |
| neutrino-8b-fv5.gguf | 4.09 GB | Fermion llama.cpp fork |
| bin/fermion-run-* | about 2 MB each | Optimized native runtime |
| MANIFEST.json and SHA sidecars | small | Integrity checks |

## Neutrino-1 0.6B

A 328 MB small model and the canonical `--draft auto` target. It can also run independently through chat, generate, serve, the native runners, or the model repository’s MLX pack.

- Runnable container: 328 MB
- Coded transport: 238 MB
- GGUF: 343 MB

## Neutrino-1 0.6B-Chat

A 328 MB conversational small model. It ships native runners and MLX code. It does not currently ship GGUF weights and should not be selected for reliable tool calling.

## Phonon speech models

The Phonon-1, Phonon-1 Micro, and Phonon-1-Big speech repositories are covered in [Speech models](https://www.fermionresearch.com/docs/speech-models/).

## License and base model

The package and model releases use Apache-2.0. Neutrino-1 8B is derived from Qwen3-8B, and the Phonon-1 models from Qwen3-ASR-0.6B; both bases use the same license.

Source: [https://www.fermionresearch.com/docs/model-reference/](https://www.fermionresearch.com/docs/model-reference/)