NVIDIA introduced Nemotron-Labs-3-Puzzle-75B-A9B, a compressed variant of Nemotron-3-Super, reducing total parameters from 120.7B to 75.3B while maintaining the same 88-block hybrid Mamba-Transformer MoE architecture. The compression process involved an iterative method of incremental compression, distillation, and re-scoring to preserve performance. Despite the reduction, the model achieves 2.03x server throughput at matched user throughput compared to the original. The release highlights advancements in efficient large language model optimization while retaining structural integrity.