Model Intelligence Sheet
1bit-MONSTER/Qwen3-Coder-30B-A3B-Instruct-GGUF overview
Qwen3 Coder 30B A3B Instruct — GGUF Q4 K M Unsloth's Q4 K M GGUF for Qwen3 Coder 30B A3B Instruct a 30B total, ~3B active coding MoE , re hosted with measured …
Runs locally from ~17.28 GB disk (24 GB VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
1 GGUF files detected
Direct downloads for local inference
| File | Type | Quantization | Size | Link |
|---|---|---|---|---|
| Qwen3-Coder-30B-A3B-Instruct-Q4_K_M.gguf | GGUF | Q4_K_M | 17.28 GB | Download |
Model Details
| Model ID | 1bit-MONSTER/Qwen3-Coder-30B-A3B-Instruct-GGUF |
|---|---|
| Author | 1bit-MONSTER |
| Pipeline | — |
| License | apache-2.0 |
| Base model | Qwen/Qwen3-Coder-30B-A3B-Instruct,unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF |
| Last modified | 2026-09-26T13:49:37.000Z |
Model README
---
license: apache-2.0
base_model:
- Qwen/Qwen3-Coder-30B-A3B-Instruct
- unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF
tags:
- gguf
- moe
- coder
---
Qwen3-Coder-30B-A3B-Instruct — GGUF (Q4_K_M)
Unsloth's Q4_K_M GGUF for Qwen3-Coder-30B-A3B-Instruct (a 30B-total, ~3B-active coding
MoE), re-hosted with measured performance for the
1bit engine on Strix Halo.
Contents
Qwen3-Coder-30B-A3B-Instruct-Q4_K_M.gguf
Measured performance (Strix Halo, Vulkan)
pp512: 1321 tok/s · tg128: 76.8 tok/s
Running it
1bit serve -m Qwen3-Coder-30B-A3B-Instruct-Q4_K_M.gguf --device vulkanAttribution
- Base model: Qwen/Qwen3-Coder-30B-A3B-Instruct, Apache 2.0.
- Quantization: unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF.
- License: Apache 2.0, inherited from the base model.
Run 1bit-MONSTER/Qwen3-Coder-30B-A3B-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