---
title: README
emoji: ⚡
colorFrom: yellow
colorTo: purple
sdk: static
pinned: false
thumbnail: >-
  https://cdn-uploads.huggingface.co/production/uploads/69e33f87d78d475613a7bd9a/l1H9DLocxXupi_GCaYEGM.png
---
<div align="center">
  <img width="700px" src="./OptAIxHuggingFace.png">
</div>

# Welcome to Opt.Gear

This is the official Hugging Face organization for **Opt.Gear**, the family of On-Device language models developed by **OptAI Inc.**
We will continuously release and update **On-Device LLM, LMM and Small Agent Model** designed for edge AI applications.

Build and scale yout AI services with models optimized specifically for On-Device deployment.

## Opt.Gear Family of Open-Weight Models
- **Opt.Gear**: A family of large language models optimized for on-device inference, available in **1B** and **270M** parameter sizes.
- **Opt.Gear-QAT**: Quantization-aware large language models designed for efficient deployment on on-device platforms.
- **Opt.Gear-1M**: An ultra-lightweight language model designed for deployment on MCUs, including platforms such as FPGA and Raspberry Pi.
- (TBU) **Opt.FGear**: A tool-calling language model designed for on-device AI agents.
- (TBU) **Opt.VGear**: A lightweight vision-language model (VLM) for a wide range of multimodal tasks.
- (TBU) **Opt.Gear-MTP**: A draft model for accelerating inference in on-device environments through multi-token prediction.