AWS M7a instance types

M7a is a general purpose Amazon EC2 family with 12 instance types, from 1 – 192 vCPUs and 4 GiB – 768 GiB of memory, running on AMD EPYC 9R14 Processor. Linux on demand starts at $0.058 per hour.

general purposex86-64
sizes
12
vCPUs
1 – 192
memory
4 GiB – 768 GiB
processor
AMD EPYC 9R14 Processor
on-demand from
$0.058

per hour · 10 regions

M7a sizes and prices

Linux, hourly, US East (N. Virginia) (other regions on each size's page)

SizevCPUMemoryLocal storageOn-demandMonthlySpot1-yr
m7a.medium14 GiB–$0.058$42.31$0.0229$0.0383
m7a.large28 GiB–$0.1159$84.62$0.048$0.0767
m7a.xlarge416 GiB–$0.2318$169$0.1103$0.1534
m7a.2xlarge832 GiB–$0.4637$338$0.2065$0.3067
m7a.4xlarge1664 GiB–$0.9274$677$0.4113$0.6135
m7a.8xlarge32128 GiB–$1.855$1,354$0.794$1.227
m7a.12xlarge48192 GiB–$2.782$2,031$1.16$1.84
m7a.16xlarge64256 GiB–$3.709$2,708$1.482$2.454
m7a.24xlarge96384 GiB–$5.564$4,062$2.076$3.681
m7a.32xlarge128512 GiB–$7.419$5,416$3.044$4.908
m7a.48xlarge192768 GiB–$11.13$8,124$4.556$7.361
m7a.metal-48xl192768 GiB–$11.13$8,124$4.954$7.361

M7a equivalents on other clouds

closest match to m7a.8xlarge (32 vCPU, 128 GiB)

AzureD32as v5Dasv5 family · 32 vCPU · 128 GiB$1.376/hr · East USDasv5 family →
Google Cloude2-standard-32E2 family · 32 vCPU · 128 GiB$1.072/hr · IowaE2 family →
AkamaiG8 Dedicated 128x32Dedicated CPU (G8) family · 32 vCPU · 128 GiB$1.68/hr · Newark, NJDedicated CPU (G8) family →
Vultrvx1-g-32c-128gVX1: General Purpose family · 32 vCPU · 128 GiB$0.96/hr · New JerseyVX1: General Purpose family →

Other general purpose families on AWS

M7a FAQ

What sizes does the AWS M7a family have?

M7a has 12 instance types, from m7a.medium (1 vCPU, 4 GiB) to m7a.metal-48xl (192 vCPU, 768 GiB). They run on AMD EPYC 9R14 Processor.

What is the cheapest M7a instance type?

m7a.medium at $0.058 per hour (Linux, on demand) in US East (N. Virginia), about $42.31 per month.

What is M7a best for?

M7a is a general purpose family. Balanced CPU and memory (about 4 GiB per vCPU) for web apps, services and most workloads.