Learn about NVFP4, a 4-bit floating-point format for high-performance inference on modern GPUs that can deliver near-baseline accuracy at large scale.

Rhdev is a blog and resource hub dedicated to Ruby on Rails development, a popular web application framework written in Ruby. Developers can explore tutorials, best practices, and case studies for building web applications with Ruby on Rails. Additionally, Rhdev covers topics such as ActiveRecord ORM, RESTful APIs, and frontend integration using JavaScript frameworks, offering insights for both beginners and experienced Rails developers.

Red Hat Developer

NVFP4 is NVIDIA's 4-bit floating-point quantization format that enables efficient deployment of large language models with near-baseline accuracy. Red Hat has released NVFP4-quantized versions of popular models ranging from 8B to 400B+ parameters, achieving 99% accuracy recovery on large models (70B-235B), 97-99% on mid-size models (~30B), and 95-98% on smaller models (7B-14B). The format provides 1.5-1.8× smaller storage than FP8 and ~3× smaller than FP16, with native hardware acceleration on NVIDIA Blackwell GPUs. Accuracy recovery improves with model size, making NVFP4 particularly effective for frontier-scale and mixture-of-experts architectures.

Accelerating large language models with NVFP4 quantization