g6.8xlargevsg7.8xlarge

g6.8xlarge is 51% cheaper than g7.8xlarge on demand in US East (N. Virginia), saving $1,512.41 a month.

At a glanceg6.8xlargeg7.8xlarge
On-demand$2.014/hr51% less$4.086/hr
Lowest rate$0.9023/hrSpot, can be interrupted$0.7699/hr15% lessSpot, can be interrupted
Size32 vCPUs, 128 GiB32 vCPUs, 128 GiB
ProcessorAMD EPYC 7R13 ProcessorIntel Xeon Scalable (Granite Rapids)
Architecturex86_64x86_64

Prices in US East (N. Virginia)

“Off” is the discount against each type’s own on-demand price. Linux, USD per hour.

Optiong6.8xlargeg7.8xlargeg7.8xlarge vs g6.8xlarge
On-demandPay by the hour$2.014$4.086▲ 103%
SpotCan be interrupted55% off$0.902381% off$0.7699▼ 15%
1-yr reservedNo upfront35% off$1.311——
3-yr reservedNo upfront54% off$0.9246——
1-yr savings planHourly commitment19% off$1.63921% off$3.218▲ 96%
3-yr savings planHourly commitment43% off$1.15646% off$2.207▲ 91%

Key differences

The biggest lead each one has.

Where g6.8xlarge leads
On-demand price$2.014 vs $4.086 per hour
51% cheaper
Where g7.8xlarge leads
Network baseline100 Gbps vs 25 Gbps
4× higher
Physical cores32 vs 16
2× higher
GPU memory32 GB vs 24 GB
33% higher

Specs

18 of 44 specs differ. Those are in bold.

Compute7 of 10 differg6.8xlargeg7.8xlarge
vCPUs3232
Physical cores1632
Threads per core21
Architecturex86_64x86_64
ProcessorAMD EPYC 7R13 ProcessorIntel Xeon Scalable (Granite Rapids)
CPU makerAMDIntel
Clock speed2.6 GHz3.9 GHz
Sustained clock3.4 GHz3.9 GHz
Processor featuresAMD Turbo; AVX; AVX2Intel AMX; Intel AVX; Intel AVX2; Intel AVX512; Intel Turbo
BurstableNoNo
MemoryAll 1 matchg6.8xlargeg7.8xlarge
Memory (GiB)128128
Accelerators2 of 3 differg6.8xlargeg7.8xlarge
GPUs11
GPU1 × NVIDIA L41 × NVIDIA RTX PRO 4500
GPU memory24 GB32 GB
Networking5 of 8 differg6.8xlargeg7.8xlarge
Network25 Gigabit100 Gigabit
Network baseline25 Gbps100 Gbps
Network peak25 Gbps100 Gbps
Enhanced networkingYes (ENA)Yes (ENA)
EFAYesYes
IPv6YesYes
Max network interfaces810
IPv4 addresses per interface3064
Storage2 of 12 differg6.8xlargeg7.8xlarge
Storage2 x 450 GB NVMe SSD1x930 GB NVMe SSD
Local disks2 × 450 GB SSD1 × 930 GB SSD
EBS optimizedYesYes
EBS over NVMeRequiredRequired
EBS throughput16000 Mbps16000 Mbps
EBS baseline bandwidth16,000 Mbps16,000 Mbps
EBS max bandwidth16,000 Mbps16,000 Mbps
EBS baseline throughput2,000 MB/s2,000 MB/s
EBS max throughput2,000 MB/s2,000 MB/s
EBS baseline IOPS64,00064,000
EBS max IOPS64,00064,000
Max EBS volumes3232
Instance details2 of 10 differg6.8xlargeg7.8xlarge
Familyg6g7
CategoryGPU instanceGPU instance
Current generationYesYes
Normalization factor2.5029821071.62152381
Bare metalNoNo
HypervisorNitroNitro
Placement groupsCluster, Partition, SpreadCluster, Partition, Spread
HibernationNoNo
Nitro EnclavesYesYes
Boot modesLegacy BIOS, UEFILegacy BIOS, UEFI

On-demand prices by region

Cheapest for g6.8xlarge first. Linux, USD per hour.

Regiong6.8xlargeg7.8xlargeDifference
US East (N. Virginia)Used in summaryus-east-1$2.014$4.086▲ 103%
US East (Ohio)us-east-2$2.014$4.086▲ 103%
US West (Oregon)us-west-2$2.014$4.086▲ 103%
Europe (Spain)eu-south-2$2.123$4.291▲ 102%
EU (Stockholm)eu-north-1$2.137$4.334▲ 103%
Asia Pacific (Mumbai)ap-south-1$2.419$4.907▲ 103%
EU (Frankfurt)eu-central-1$2.519$5.11▲ 103%
EU (London)eu-west-2$2.557$6.198▲ 142%
Asia Pacific (Sydney)ap-southeast-2$2.619$5.313▲ 103%
South America (Sao Paulo)sa-east-1$3.424$8.172▲ 139%

Keep exploring

Each type’s own page, and how it compares with others.

g6.8xlargeGPU instance, g6 family
g6.8xlarge prices and specsEvery size in the g6 family

Compare g6.8xlarge with

g6e.8xlargeg5.8xlarge
g7.8xlargeGPU instance, g7 family
g7.8xlarge prices and specsEvery size in the g7 family

Compare g7.8xlarge with

g7e.8xlarge

Pay the committed rate, with commitment protection

The best committed rate on these two is 46–54% below on-demand. Connect your cloud accounts to unlock greater savings with commitment protection.

Connect AWS account

AWS EC2 Linux prices in USD. Monthly figures use 730 hours. Prices last updated 12 Nov 2026.