Flopper.io Flopper.io
GPUs Systems Datacenters Desktops Providers Pricing
  • Home
  • Glossary
  • precision

BF16

precision
Also known as: BFloat16, Brain Float

Definition

Brain Float 16 - A 16-bit floating-point format developed by Google with the same exponent range as FP32 but reduced mantissa precision. Provides better numerical stability than FP16 for training large language models while maintaining memory and performance benefits.

GPUs featuring BF16

Huawei Ascend 970

Da Vinci v5 · 2028 · 288 GB

Huawei Ascend 960

Da Vinci v4 (chiplet) · 2027 · 288 GB

NVIDIA Rubin

Rubin · 2026 · 288 GB

NVIDIA Vera Rubin

Rubin · 2026 · 576 GB

Huawei Ascend 950DT

Da Vinci v3 · 2026 · 144 GB

Huawei Ascend 950PR

Da Vinci v3 · 2026 · 128 GB
Browse All GPUs →

Related Terms

FP16

16-bit floating-point precision format (half precision). Commonly used in AI tra...

TF32

TensorFloat-32 - A 19-bit precision format introduced with NVIDIA Ampere archite...

FP8

8-bit floating-point precision format introduced in Hopper architecture. Enables...

More precision Terms

TF32

TensorFloat-32 - A 19-bit precision format introduced with NVIDIA Ampere archite...

FP16

16-bit floating-point precision format (half precision). Commonly used in AI tra...

FP8

8-bit floating-point precision format introduced in Hopper architecture. Enables...

Explore More Terms

Browse the full glossary or dive into architecture guides

Back to Glossary Architecture Guides
Flopper.io Flopper.io

Compare the GPUs powering AI infrastructure.

Add your company

Stay updated

New GPU data, benchmarks, and pricing changes delivered weekly.

Company
About Flopper Contact How to Use Submit GPU Specs
Free Tools
FLOPS Calculator GPU Cloud Pricing Compare GPUs AI Power Translator
My other projects
boulderinglist.com llmstxt.studio paletter.dev probe.bike radiusing.uk rides.bike
Popular GPUs
NVIDIA H100 NVIDIA H200 NVIDIA A100 NVIDIA B200
All Systems AI Datacenters Compare Systems Compare

© 2026 Flopper.io - AI GPU Performance Benchmarks

Free tools for AI infrastructure planning

Stay Updated on GPU Specs

Get weekly updates on the latest datacenter GPUs, benchmark results, and performance metrics.