KingNish/Reasoning-0.5b

KingNish/Reasoning-0.5b is a 0.5 billion parameter Qwen2.5-based instruction-tuned causal language model developed by Nishith Jain, specifically designed for explicit reasoning tasks. This model is trained to first generate a separate reasoning step before producing its final answer, making its thought process transparent. It excels in scenarios requiring structured problem-solving and supports a 32768-token context length across multiple languages.