liodon-ai/Qwen3-4B-Thinking-2507-imatrix-GGUF overview
Qwen3 4B Thinking 2507 — iMatrix GGUF GGUF quantizations of Qwen/Qwen3 4B Thinking 2507 https://huggingface.co/Qwen/Qwen3 4B Thinking 2507 , published by Liodo…
Runs locally from ~1.41 GB disk (4 GB VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
| File | Type | Quantization | Size | Link |
|---|---|---|---|---|
| Qwen3-4B-Thinking-2507-IQ2_M.gguf | GGUF | IQ2_M | 1.41 GB | Download |
| Qwen3-4B-Thinking-2507-IQ3_M.gguf | GGUF | IQ3_M | 1.83 GB | Download |
| Qwen3-4B-Thinking-2507-IQ4_XS.gguf | GGUF | IQ4_XS | 2.11 GB | Download |
| Qwen3-4B-Thinking-2507-Q4_K_M.gguf | GGUF | Q4_K_M | 2.33 GB | Download |
| Qwen3-4B-Thinking-2507-Q5_K_M.gguf | GGUF | Q5_K_M | 2.69 GB | Download |
| Qwen3-4B-Thinking-2507-Q6_K.gguf | GGUF | Q6_K | 3.08 GB | Download |
| Qwen3-4B-Thinking-2507-Q8_0.gguf | GGUF | Q8_0 | 3.99 GB | Download |
Model Details
| Model ID | liodon-ai/Qwen3-4B-Thinking-2507-imatrix-GGUF |
|---|---|
| Author | liodon-ai |
| Pipeline | text-generation |
| License | other |
| Base model | Qwen/Qwen3-4B-Thinking-2507 |
| Last modified | 2026-07-13T04:55:19.000Z |
Model README
---
license: other
base_model: Qwen/Qwen3-4B-Thinking-2507
base_model_relation: quantized
pipeline_tag: text-generation
library_name: gguf
tags:
- gguf
- ollama
- local-llm
- llama.cpp
- lm-studio
- quantized
- imatrix
- sub-4-bit
- qwen3
quantized_by: liodon-ai
---
Qwen3-4B-Thinking-2507 — iMatrix GGUF
GGUF quantizations of Qwen/Qwen3-4B-Thinking-2507, published by Liodon AI.
Quick Start
llama.cpp
llama-cli -hf liodon-ai/Qwen3-4B-Thinking-2507-imatrix-GGUF:Q4_K_MOllama
ollama run hf.co/liodon-ai/Qwen3-4B-Thinking-2507-imatrix-GGUF:Q4_K_MLM Studio / Jan — search liodon-ai/Qwen3-4B-Thinking-2507-imatrix-GGUF and pick your quant.
Quants
| Quant | Size | VRAM est. | Notes |
|-------|------|-----------|-------|
| IQ2_M | 1.51 GB | ~2 GB | 2-bit, iMatrix — smallest usable |
| IQ3_M | 1.96 GB | ~2 GB | 3-bit, iMatrix — great quality/size tradeoff |
| IQ4_XS | 2.27 GB | ~3 GB | 4-bit extra-small, iMatrix |
| Q4_K_M | 2.50 GB | ~3 GB | 4-bit, iMatrix-calibrated (recommended) |
| Q5_K_M | 2.89 GB | ~3 GB | 5-bit, iMatrix-calibrated |
| Q6_K | 3.31 GB | ~4 GB | 6-bit, iMatrix-calibrated, near-lossless |
| Q8_0 | 4.28 GB | ~5 GB | 8-bit, essentially lossless |
What is iMatrix?
Standard quantization treats all weights equally. iMatrix runs 128 calibration chunks through
the full-precision model to find which weights matter most, then allocates more precision where
it counts. At Q2/Q3/Q4 this means noticeably better coherence and instruction-following —
same file size, better output.
Calibration: 2M tokens of WikiText-103.
> Also see plain (non-iMatrix) quants: liodon-ai/Qwen3-4B-Thinking-2507-GGUF
Source
- Model: Qwen/Qwen3-4B-Thinking-2507
- License: other
---
Quantized by Liodon AI
Run liodon-ai/Qwen3-4B-Thinking-2507-imatrix-GGUF with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.
Source: Hugging Face · Compare models