Jackrong/gpt-oss-120b-Distill-Qwen3-4B-Thinking
Jackrong/gpt-oss-120b-Distill-Qwen3-4B-Thinking is a 4-billion parameter language model, distilled from the gpt-oss-120b-high model and built on the Qwen3-4B architecture. It is specifically optimized for human-friendly, high-fidelity reasoning, featuring an explicit point-by-point thought chain. With a maximum context length of 32,768 tokens, this model excels at complex analytical tasks, technical tutorials, and user education requiring transparent, multi-step logic.
Loading preview...
Model Overview
Jackrong/gpt-oss-120b-Distill-Qwen3-4B-Thinking is a specialized 4-billion parameter language model, developed by Jackrong, the gpt-oss team, and Qwen authors. It is a deeply distilled and fine-tuned variant of the gpt-oss-120b-high model, utilizing the lightweight Qwen3-4B architecture. The model is designed to preserve the complex multi-step reasoning patterns of its larger source model while operating efficiently at a smaller scale.
Key Capabilities
- High-Fidelity Reasoning: Optimized for human-friendly, complex reasoning tasks.
- Transparent Thought Chains: Generates explicit, point-by-point thought chains (e.g., bullet-point steps) to make intricate logic clear and easy to follow.
- Extended Context Window: Supports a maximum context length of 32,768 tokens, enabling long-form reasoning without truncation.
- Efficient Distillation: Compresses advanced reasoning capabilities onto a 4B-parameter backbone, making it more accessible.
Recommended Use Cases
- Technical Tutorials: Ideal for generating stepwise code walkthroughs and explanations.
- Complex Queries: Suitable for math, engineering, and other fields requiring deep reasoning to avoid oversimplified answers.
- User Education: Provides clear, scannable outputs that aid learning and reduce confusion.
- Moderation and Analysis: Its structured output format facilitates programmatic parsing of responses.