Hugging Face TinyR1-32B-Preview is a 32 billion parameter reasoning model developed by qihoo360, based on the Deepseek-R1-Distill-Qwen-32B architecture. It is specifically optimized for complex reasoning tasks across mathematics, coding, and science domains, demonstrating performance in math that nearly matches larger models. This model was created by fine-tuning and merging domain-specific models to achieve strong overall performance in these analytical areas.
No reviews yet. Be the first to review!