AWS Amazon EC2

g4dn.12xlarge

⚡ 4x GPU • 64 GB

Part of the G4DN Accelerated Computing family — Hardware accelerators (GPUs, FPGAs) for specialized compute.

Name:gGPU Graphics & ML Inference44th GenerationdDirect-Attached NVMe SSDnEnhanced Networking12xlarge12 Extra Large

On-demand, Linux

$3.9120/ hr

~$2,855.76/mo · lowest in US East (N. Virginia), us-east-1

Savings Plan, Linux

$2.0280/ hr

From, with a 3-year Compute Savings Plan, No Upfront: −48% vs on-demand rate (us-east-1)

Per unit

$0.0815/ vCPU-hr

$0.0204/GB-hr

Compute

48vCPUs

24 cores • x86_64

Memory

192GB

4.0 GB / vCPU

Network

50 Gbps

Baseline: 50 Gbps

Storage & EBS

1,187.5MB/s

Max: 40,000 IOPS

Alternatives to g4dn.12xlarge

Option	Price vs g4dn.12xlarge	Why look at it
g5.12xlarge Compare →	+45%+$1.76/hr Linux on-demand, us-east-1	AMD (x86_64), newer generation, same vCPU and RAM
g6.12xlarge Compare →	+18%+$0.69/hr Linux on-demand, us-east-1	AMD (x86_64), newer generation, same vCPU and RAM

On Google Cloud

g4dn.12xlarge is an accelerated machine. Those are compared by the accelerator, not by vCPU and memory, so this page doesn't pick a match. Browse Google Cloud machines.

Technical Specifications & Limits

Hardware specs, container thresholds, and system limits

Compute & Processor

ProcessorIntel Xeon Family

Architecturex86_64 (x86_64)

vCPUs / Threads48 vCPUs

Physical Cores24

Clock Speed2.5 GHz

Intel AVX / AVX2AVX & AVX2 Supported

Intel Turbo BoostYes

Bare MetalNo (Virtualized)

Performance & Benchmarks

CoreMark Score715,777

CoreMark / vCPU14,912 / core

FFmpeg Transcoding110 FPS

CoreMark per Dollar182,970 score / $

Cost per vCPU-hr$0.0815

Cost per GB-hr$0.0204

Containers & Kubernetes

EKS Max Pods234 pods

Max ENIs8

IPv4/v6 per ENI30

Total Addressable IPs240 total IPs

CNI Prefix DelegationSupported (High Density)

Networking & Bandwidth

Network Performance50 Gbps

Baseline Bandwidth50 Gbps

Burst / Peak Bandwidth50 Gbps

Enhanced NetworkingYes

EFA SupportNo

IPv6 SupportNative Dual-Stack Supported

Storage & EBS Optimization

Storage ConfigurationNVMe SSD (900 GB)

Local NVMe Disk900 GB NVMe SSD

EBS Baseline Throughput1187.5 MB/s

EBS Max Throughput1187.5 MB/s

EBS Baseline IOPS40,000

EBS Max / Burst IOPS40,000

Dedicated EBS Throughput7000 Mbps

GPU & Hardware Acceleration

GPU ModelNVIDIA T4

Accelerator Count4x GPU(s)

Total GPU Memory64.0 GB

Memory Architecture—

Acceleration RuntimesCUDA, TensorRT, cuDNN, PyTorch, PyJAX

Platform & Architecture Details

Hypervisor / SystemAWS Nitro System (PCIe Offload)

Nitro EnclavesSupported (Isolated security enclaves)

Normalization Factor7.437262357

Generation StatusCurrent Generation (Recommended)

Rates by region

Live on-demand rates across all global AWS cloud regions

Lowest Price Region

$3.9120/ hr

US East (N. Virginia) (us-east-1)

Highest Price Region

$6.6490/ hr

South America (Sao Paulo) (sa-east-1)

Global Fleet Average

$4.8397/ hr

Across all active regions

Regional Coverage

26regions

Available worldwide

Linux — Available in all 26 regions
Region	On-Demand Price	% vs On-Demand Rate
af-south-1 Africa (Cape Town)	$5.1900
ap-east-1 Asia Pacific (Hong Kong)	$6.0240
ap-south-1 Asia Pacific (Mumbai)	$4.3060
ap-northeast-3 Asia Pacific (Osaka)	$5.2810
ap-northeast-2 Asia Pacific (Seoul)	$4.8120
ap-southeast-1 Asia Pacific (Singapore)	$5.4740
ap-southeast-2 Asia Pacific (Sydney)	$5.0870
ap-northeast-1 Asia Pacific (Tokyo)	$5.2810
us-gov-east-1 AWS GovCloud (US-East)	$4.9310
us-gov-west-1 AWS GovCloud (US-West)	$4.9310
ca-central-1 Canada (Central)	$4.3430
eu-central-1 EU (Frankfurt)	$4.8900
eu-west-1 EU (Ireland)	$4.3620
eu-west-2 EU (London)	$4.5770
eu-south-1 EU (Milan)	$4.5800
eu-west-3 EU (Paris)	$4.5740
eu-north-1 EU (Stockholm)	$4.1500
me-south-1 Middle East (Bahrain)	$4.7980
eu-west-3-wl1-cmn1 Morocco (Casablanca)	$5.2340
eu-west-3-wl1-dss1 Senegal (Dakar)	$5.2340
sa-east-1 South America (Sao Paulo) Highest	$6.6490
us-east-1 US East (N. Virginia) Lowest	$3.9120
us-east-2 US East (Ohio) Lowest	$3.9120
us-west-2-lax-1 US West (Los Angeles)	$4.6940
us-west-1 US West (N. California)	$4.6940
us-west-2 US West (Oregon) Lowest	$3.9120

Windows — Available in all 26 regions
Region	On-Demand Price	% vs On-Demand Rate
af-south-1 Africa (Cape Town)	$7.3980
ap-east-1 Asia Pacific (Hong Kong)	$8.2320
ap-south-1 Asia Pacific (Mumbai)	$6.5140
ap-northeast-3 Asia Pacific (Osaka)	$7.4890
ap-northeast-2 Asia Pacific (Seoul)	$7.0200
ap-southeast-1 Asia Pacific (Singapore)	$7.6820
ap-southeast-2 Asia Pacific (Sydney)	$7.2950
ap-northeast-1 Asia Pacific (Tokyo)	$7.4890
us-gov-east-1 AWS GovCloud (US-East)	$7.1390
us-gov-west-1 AWS GovCloud (US-West)	$7.1390
ca-central-1 Canada (Central)	$6.5510
eu-central-1 EU (Frankfurt)	$7.0980
eu-west-1 EU (Ireland)	$6.5700
eu-west-2 EU (London)	$6.7850
eu-south-1 EU (Milan)	$6.7880
eu-west-3 EU (Paris)	$6.7820
eu-north-1 EU (Stockholm)	$6.3580
me-south-1 Middle East (Bahrain)	$7.0060
eu-west-3-wl1-cmn1 Morocco (Casablanca)	$7.4420
eu-west-3-wl1-dss1 Senegal (Dakar)	$7.4420
sa-east-1 South America (Sao Paulo) Highest	$8.8570
us-east-1 US East (N. Virginia) Lowest	$6.1200
us-east-2 US East (Ohio) Lowest	$6.1200
us-west-2-lax-1 US West (Los Angeles)	$6.9020
us-west-1 US West (N. California)	$6.9020
us-west-2 US West (Oregon) Lowest	$6.1200

Savings Plans: what to know

A Savings Plan commits you to a dollar amount per hour of spend for one or three years, not to an instance: usage up to that amount is billed at the plan's rates and the rest at on-demand, and unused commitment is still billed.
A Compute Savings Plan follows your usage across instance families, sizes, regions and operating systems; an EC2 Instance Savings Plan is locked to one family in one region in exchange for a lower rate (what that flexibility costs).
In US East (N. Virginia) (us-east-1), the 3-year Compute Savings Plan rate (No Upfront) for g4dn.12xlarge on Linux is 52% of the on-demand rate, so the instance has to run at least 52% of the hours for the commitment to cost less than on-demand.
On Windows a plan discounts only the compute part of the rate: the Windows license part costs the same as on demand, so the percentage off is smaller than on Linux.
Rates are AWS's effective hourly rate for each payment option; All Upfront pays the whole term's commitment at purchase, which is why its hourly rate is lower; Partial Upfront (at least half up front) isn't shown on this page. The Savings Plans calculator has all three and sizes a commitment for this instance or a mix of them. How AWS applies a plan each hour: how Compute Savings Plans decide what they cover. Sizing one for elastic workloads: managing Compute Savings Plans.