NVIDIA Releases NVFP4 MiniMax-M3: 428B MoE Model at 4-bit on Blackwell

NVIDIA published an NVFP4-quantized version of MiniMax-M3, a 428B-parameter multimodal MoE model with a 1M-token context window, on Hugging Face. The 4-bit quantization targets Blackwell GPUs and delivers roughly 2x memory reduction compared to full precision, making the model more practical to deploy at scale.