1bit-MONSTER/Qwen2.5-3B-Instruct-GGUF overview
Qwen2.5 3B Instruct — GGUF Q4 K M Qwen's own Q4 K M GGUF, re hosted with measured performance for the 1bit engine https://github.com/1bit MONSTER/engine on Str…
Runs locally from ~1.96 GB disk (4 GB VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
| File | Type | Quantization | Size | Link |
|---|---|---|---|---|
| qwen2.5-3b-instruct-q4_k_m.gguf | GGUF | Q4_K_M | 1.96 GB | Download |
Model Details
| Model ID | 1bit-MONSTER/Qwen2.5-3B-Instruct-GGUF |
|---|---|
| Author | 1bit-MONSTER |
| Pipeline | — |
| License | other |
| Base model | Qwen/Qwen2.5-3B-Instruct |
| Last modified | 2026-09-26T13:53:18.000Z |
Model README
---
license: other
license_name: qwen-research
license_link: LICENSE
base_model:
- Qwen/Qwen2.5-3B-Instruct
tags:
- gguf
---
Qwen2.5-3B-Instruct — GGUF (Q4_K_M)
Qwen's own Q4_K_M GGUF, re-hosted with measured performance for the
1bit engine on Strix Halo.
Non-commercial only — unlike most Qwen2.5 sizes, the 3B checkpoint is released
under the Qwen Research License, not Apache 2.0. It's licensed for research
and evaluation use only; commercial use requires a separate license from Alibaba Cloud.
Contents
qwen2.5-3b-instruct-q4_k_m.gguf
Measured performance (Strix Halo, Vulkan)
pp512: 2995 tok/s · tg128: 91.7 tok/s
Running it
1bit serve -m qwen2.5-3b-instruct-q4_k_m.gguf --device vulkanAttribution
- Model and quantization: Qwen/Qwen2.5-3B-Instruct-GGUF.
- Qwen is licensed under the Qwen RESEARCH LICENSE AGREEMENT, Copyright (c) Alibaba Cloud. All Rights Reserved.
- Built with Qwen.
Run 1bit-MONSTER/Qwen2.5-3B-Instruct-GGUF with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.
Source: Hugging Face · Compare models