PKU-DS-LAB/FairyR1-32B
TEXT GENERATIONConcurrency Cost:2Model Size:32BQuant:FP8Ctx Length:32kPublished:May 23, 2025License:apache-2.0Architecture:Transformer0.1K Open Weights Cold

FairyR1-32B is a 32 billion parameter reasoning model developed by PKU-DS-LAB, built upon the DeepSeek-R1-Distill-Qwen-32B base with a 32768 token context length. It leverages a novel "distill-and-merge" pipeline to achieve performance comparable to much larger models in mathematical and coding tasks. This model is optimized for efficiency, offering strong task-specific performance with significantly reduced parameters and inference costs.

Loading preview...