Moxin AI Hero Image

Model Hub

Open-source models for general use, reasoning, and multimodal understanding.

Moxin-7B-Base

Foundational pre-trained model for custom fine-tuning. Enhanced Mistral architecture trained on 2+ trillion tokens.

HF Card GitHub

Moxin-7B-Instruct

AI assistant fine-tuned for instruction following and dialogue. Uses SFT and DPO with Tülu 3 framework.

HF Card GitHub

Moxin-7B-Reasoning

Advanced reasoning for math, logic, and coding. Enhanced with GRPO reinforcement learning.

HF Card GitHub

Moxin-7B-VLM

Vision-language model for multimodal understanding. Combines Moxin-7B-Base with DINOv2 and SigLIP encoders.

HF Card GitHub

Ready to Build?

Models, code, and data available on Hugging Face and GitHub.

View on Hugging Face
View on GitHub