TeichAI/Qwen3-4B-Thinking-2507-MiniMax-M2.1-Distill
The TeichAI/Qwen3-4B-Thinking-2507-MiniMax-M2.1-Distill is a 4 billion parameter Qwen3-based language model, fine-tuned on a reasoning dataset derived from MiniMax M2.1. This model is specifically optimized for complex reasoning tasks across domains like coding, science, and deep research. It leverages the unsloth/Qwen3-4B-Thinking-2507 base model and was trained efficiently using Unsloth and Huggingface's TRL library.
Loading preview...
Model Overview
TeichAI/Qwen3-4B-Thinking-2507-MiniMax-M2.1-Distill is a 4 billion parameter language model built upon the Qwen3 architecture. It has been specifically fine-tuned using a reasoning dataset sourced from MiniMax M2.1, aiming to enhance its capabilities in complex analytical and problem-solving scenarios. The training process utilized unsloth/Qwen3-4B-Thinking-2507 as its base model and was accelerated by Unsloth and Huggingface's TRL library, enabling faster iteration and development.
Key Capabilities
This model is designed with a strong focus on reasoning, making it particularly suitable for:
- Coding: Generating and understanding code, potentially assisting in debugging or development.
- Science: Processing scientific texts, aiding in research, and understanding complex scientific concepts.
- Deep Research: Facilitating in-depth analysis and synthesis of information across various domains.
Training Details
The model's training involved a dedicated reasoning dataset, TeichAI/MiniMax-M2.1-8800x, incurring a cost of approximately $42.94 USD and processing a total of 39.2 million input and output tokens. This targeted training on a specialized reasoning dataset is intended to imbue the model with enhanced logical inference and analytical skills.