NVIDIA Model Optimizer (formerly TensorRT Model Optimizer) is a unified library of state-of-the-art model optimization techniques including quantization, pruning, distillation, speculative decoding, and sparsity. It compresses deep learning models for downstream deployment frameworks like TensorRT-LLM, TensorRT, and vLLM to dramatically optimize inference speed. The library supports highly performant quantization formats including NVFP4, FP8, INT8, and INT4 with advanced algorithms such as SmoothQuant, AWQ, and SVDQuant.