Qwen3-4B GGUF Text Encoder

GGUF quantized version of Qwen3-4B language model, essential text encoder for Z-Image GGUF deployments with thinking mode support

Text Encoder

LLM

GGUF

Qwen3

4B Parameters

Bilingual

Thinking Mode

Visit Resource

Overview

Qwen3-4B-GGUF is the required text encoder for Z-Image GGUF deployments, providing bilingual (Chinese/English) understanding and advanced reasoning capabilities. Supports unique thinking mode for complex logical reasoning tasks.

Features

4B parameters with efficient GGUF quantization
Seamless switching between thinking and non-thinking modes
Excellent bilingual text understanding (Chinese & English)
32K native context length (131K with YaRN)
Superior reasoning, coding, and math capabilities
36 layers with GQA architecture (32 Q heads, 8 KV heads)
Multiple quantization levels (Q4_K_M recommended)

Installation

Download GGUF file and place in ComfyUI/models/text_encoders/ directory

Usage

Use as text encoder with Z-Image GGUF models in ComfyUI. Supports enable_thinking parameter for activating reasoning mode.

Requirements

4GB+ VRAM for text encoding
Compatible with Z-Image GGUF models
ComfyUI GGUF extension