Quran-R1: Reasoning on Quranic Content

Quran-R1 is a specialized 0.8 billion parameter language model, fine-tuned from Qwen3-0.6B by khazarai. Its primary purpose is to perform reasoning and question-answering tasks specifically on the Quran, providing detailed, structured reasoning steps along with its answers. This model was developed using supervised fine-tuning (SFT) on the musaoc/Quran-reasoning-SFT dataset, which emphasizes complex, knowledge-intensive questions about Islamic scripture.

Key Capabilities

• Structured Reasoning: Generates answers accompanied by explicit reasoning traces.
• Quranic QA: Excels at answering questions related to the Quran's content.
• Educational Support: Useful for assisting with structured understanding of Quranic topics.
• Research Tool: Provides a platform for exploring reasoning capabilities of smaller LLMs on religious texts.

Good for

• Educational applications focused on Quranic studies.
• Research into fine-tuning LLMs for domain-specific reasoning.
• Developing QA systems that require reasoned responses on religious texts.

Limitations

It is important to note that Quran-R1 is not intended for authoritative religious rulings (fatwas) or sensitive theological debates. Its outputs reflect dataset biases and may not cover all scholarly interpretations. Users should be aware of the risk of hallucination, common to all LLMs, and use the model with caution in sensitive religious contexts.