README.md

---
title: README
emoji: 🦀
colorFrom: pink
colorTo: purple
sdk: static
pinned: false
license: mit
thumbnail: >-
  https://cdn-uploads.huggingface.co/production/uploads/684e077edcb41cdac6183e2d/cFE0tIZqexfAFbsgnWq-U.png
---
<div align="center">
  <img src="cq_black.png"/>
  <h1>🚀 CQ Systems</h1>
  <h3><i>To Support The Democratization of Foundational Intelligence</i></h3>
  <a href="https://www.cqurity.ai" target="_blank">www.cqurity.ai</a>
</div>

---

## 🌿 Our Mission: Sovereignty & Sustainability

At **CQ Systems**, we believe that the future of intelligence must be decentralized. The current trajectory of AI development relies heavily on massive, energy-intensive data centers that centralize power and compromise privacy. We are working to break this cycle through two core pillars:

### 🌍 Ecological Responsibility
By optimizing state-of-the-art models for consumer hardware, edge devices, and advanced silicon architectures, we aim to reduce the global reliance on hyper-scale data centers. Moving computation from massive server farms to local, efficient execution minimizes the carbon footprint of AI inference, promoting a more sustainable digital ecosystem.

### 🔐 Privacy & Assured Execution
We are pushing privacy and assured execution back into the hands of the user. By enabling frontier-level models to run locally, we eliminate the need to transmit sensitive data to third-party clouds. This approach significantly reduces the regulatory hurdles associated with data retention and privacy compliance, empowering individuals and organizations to maintain absolute control over their own systems and information.

**We aren't just optimizing weights; we are democratizing foundational intelligence.**

---

## 🛠 Our Work
We specialize in highly optimized **GGUF quantizations** of top-tier foundation models. Our current focus is on the Google Gemma 4 family, providing a spectrum of models that balance extreme memory efficiency with uncompromising reasoning capabilities.

### 🌟 Model Repositories

| Model Repository | Size / Type | Use Case |
| :--- | :--- | :--- |
| [**Google-Gemma-4-31B-It-GGUF**](https://huggingface.co/CQSystems/Google-Gemma-4-31B-It-GGUF) | 31B Dense | The heavyweight flagship; compressed for local execution without sacrificing frontier reasoning. |
| [**CQ-Gemma-4-26B-A4B-It-GGUF**](https://huggingface.co/CQSystems/CQ-Gemma-4-26B-A4B-It-GGUF) | 26B (Active 4B) | A precision-engineered balance of depth and efficiency for high-performance local tasks. |
| [**Google-Gemma-4-E4B-It-GGUF**](https://huggingface.co/CQSystems/Google-Gemma-4-E4B-It-GGUF) | 8B (4.5B Active) | *Our most popular release.* Optimized via iMatrix for Apple Silicon and edge devices. |
| [**Google-Gemma-4-E2B-It-GGUF**](https://huggingface.co/CQSystems/Google-Gemma-4-E2B-It-GGUF) | 5B (2.3B Active) | Ultra-lightweight and lightning fast; purpose-built for constrained memory environments. |

---

## 🤝 Connect with Us
* 🌐 **Website:** [cqurity.ai](https://www.cqurity.ai)
* 💡 **Interests:** AI, Machine Learning, Quantization, Local LLMs, Model Optimization, Edge Computing

*Building the future of local AI, one optimized tensor at a time.*