Open Format · Deep Learning · 8-Bit/4-Bit Precision

HiFloat: Advanced AI low bit-width floating-point format

HiFloat (HiF8/HiF4/...) is an efficient floating-point format for AI computing. It outperforms FP16/BF16 and FP8 by delivering higher computing efficiency under identical hardware constraints.
Experimental results for quantization of Gaussian-distributed random numbers, evaluated by MSE and SQNR
quantization_error_reduction
−24%
Assessment of chip PPA gains based on the assumption that the computing power of the HiF4 matrix is twice that of the 8-bit (FP8/INT8) format
incrementa_chip_area_savings
68%
power_reduction
10%

Hardware Support

Events and Meetings

Community News

CNCC Workshop

Advancing Low-Precision Computing for Efficient Large-Model Deployment: GCC-HiFloat Community Hosts Workshop at CNCC 2026 in Chengdu

On October 22, 2026, the GCC-HiFloat community hosted a low-precision computing workshop at CNCC 2026. Experts from universities, research institutes, and industry shared research findings and practical experience in numerical principles, data formats, model quantization, and hardware–software co-optimization, exploring ways to improve the efficiency of large-model training and inference.

2026/10/26
CNCC Workshop

Low-Precision Computing at CNCC: New Frontiers in Data Formats, Algorithms, and System Co-Design

In the era of large models, how can fewer bits deliver greater computational efficiency? On October 22, the GCC-HiFloat community will host the Low-Precision Computing Forum at CNCC 2026. The forum will focus on frontier research in low-precision data formats, model training and inference, quantization algorithms, hardware–software co-design, and system optimization, examining key challenges and recent advances from numerical representation to efficient AI computing systems.

2026/10/10
HiFloat AICAS IEEE

HiFloat Community Participates in IEEE AICAS 2026, Bringing Low-Bit Data Formats to an International Audience

During IEEE AICAS 2026, the GCC-HiFloat community along with Huawei, a GCC council member and platinum member, and Professor Li Du’s team from Nanjing University for in-depth discussions with university faculty, students, and industry representatives from around the world on the capabilities enabled by HiFloat and research trends in IEEE standardization.

2026/09/24
HiFloat4 Association Standard

Call for Contributors: HiFloat4 Data Format Technical Specification Opens for Participating Organizations

The HiFloat4 Data Format Technical Specification, an association standard initiated by the HiFloat community under the GCC Intelligent Computing Industry Development Committee, is scheduled for publication in February 2027. Applications are now open for contributing organizations, and GCC members are welcome to participate in drafting and discussions.

2026/09/01
HiFloat Community Launch

HiFloat Community Officially Launches, Accelerating the Low-Precision Computing Ecosystem

On August 26, the Low-Precision Computing Technology Workshop and HiFloat Community Launch Conference took place in Beijing. Covering low-bit data format standards, hardware–software co-design, HiFloat technologies and applications, Ascend parallel acceleration, efficient data transmission, and model quantization, the conference officially launched the GCC-HiFloat community and unveiled its initial achievements. Experts from research institutes, universities, chip companies, and AI enterprises gathered to explore the next steps from technical innovation to standardization and industrial adoption.

2026/8/28
ICME Grand Challenge

Low-Bit Computing Opens New Paths to Greater Large-Model Efficiency

On the afternoon of July 5, the IEEE ICME 2026 Workshop on the Low-Bit Large-Model Efficient Computing Challenge took place in Bangkok, Thailand. Focusing on frontier topics including low-bit quantization of large models and video generation model inference and training, the workshop brought together outstanding teams from universities, research institutes, and industry to share their challenge results.

2026/07/06

Building the HiFloat Ecosystem Together

25+ partners

Academia & Research

  • Peking University Shenzhen Graduate School
  • Beihang University
  • South China University of Technology
  • Peng Cheng Laboratory
  • Standardization Working Group 32 on Intelligent Computing
  • Tsinghua Shenzhen International Graduate School
  • Shanghai Artificial Intelligence Laboratory
  • Shenzhen Research Institute of Big Data

Industry Institutions

  • XFUSION
  • Phytium
  • GRG Wuzhou
  • HUAWEI
  • iFLYTEK
  • AtomGit
  • LongSailing
  • Modelers
  • Amedac
  • SenseTime
  • KUNTAI
  • TIANKUAN
  • CTYun
  • LECARC
  • YSEMI
  • China Telecom Research Institute
  • X Core Sigma
Note: Partners are listed in alphabetical order.

Subscribe to the HiFloat Mailing List