How to Install and Run Huihui-AI Models: A Complete Step-by-Step Guide

If you have been exploring the world of open-weight large language models, you have likely come across huihui-ai on platforms like Hugging Face. Known popularly for publishing high-performance, “abliterated” (refusal-removed/uncensored) variants of mainstream models like Qwen, Llama, Gemma, and DeepSeek, huihui-ai does not host an independent consumer chat website. Instead, they provide specialized weight modifications that users can download and run locally.

If you are wondering how to install (“как установить”) and run these models on your own machine, this guide covers everything you need to know.

Understanding What You Are Installing

Before jumping into the installation, it helps to clear up a common point of confusion: Huihui-ai is a model creator/publisher, not a standalone software application.

To use a huihui-ai model (such as an abliterated Qwen or DeepSeek variant), you need a local AI runner or backend client. The two most popular and beginner-friendly ways to install and run them are via Ollama (for command-line or lightweight use) or LM Studio / Pinokio (for graphical user interfaces).

Method 1: Installing via Ollama (Easiest & Fastest)

Ollama allows you to download and run huihui-ai model weights directly through your terminal using simple commands.

Step 1: Download and Install Ollama

1. Go to the official Ollama website.
2. Download the installer for your operating system (Windows, macOS, or Linux).
3. Run the installer and follow the standard on-screen setup prompts.

Step 2: Open Your Terminal

• Windows: Open Command Prompt (cmd) or PowerShell.
• macOS / Linux: Open the Terminal app.

Step 3: Run the Huihui-AI Model Command

Many popular huihui-ai builds are packaged for community runners. For example, to pull and run an abliterated model variant, you can use community-mapped tags or pull directly from repositories.

Type the following command into your terminal (using a popular variant like an abliterated DeepSeek or Qwen build as an example):

ollama run huihui_ai/deepseek-r1-abliterated:8b

(Note: You can swap out the model tag depending on which exact variant or size you prefer from the community libraries).

Ollama will automatically download the necessary GGUF weight files and launch an interactive chat right inside your command line window.

Method 2: Installing via LM Studio (Best Graphical Interface)

If you prefer a ChatGPT-style desktop application with a clean UI rather than using a terminal, LM Studio is the ideal choice. It allows you to download GGUF files directly from Hugging Face, where huihui-ai hosts its models.

Step 1: Install LM Studio

1. Visit the official LM Studio website.
2. Download and install the application for Windows, macOS, or Linux.

Step 2: Search for Huihui-AI Models

1. Open LM Studio after installation.
2. Use the Search bar at the top of the application.
3. Type huihui-ai or the specific model name you want (