instancefinder
InstanceFinder / GCP / a2-highgpu-8g

a2-highgpu-8g

GCP machine type · 96 vCPU · 680 GiB RAM · x86_64

On-demand
$5.9158/hr
Monthly
$4,318.55/mo
vCPU
96
Memory
680 GiB
Comparable 96 vCPU / 680 GiB instances on other clouds: AWS · Azure · OCI
See a2-highgpu-8g in the full interactive table →

The a2-highgpu-8g is a GCP machine type with 96 vCPUs and 680 GiB of memory (7.083 GiB per vCPU), on the x86_64 architecture. On-demand pricing starts at $5.9158/hr ($4,318.55/mo) in us-central1. It is available in 10 GCP region(s). Not affiliated with Google.

Pricing — on-demand, us-central1

Hourly (USD)$5.9158
Monthly (USD, 730 h)$4,318.55
Per vCPU-hour (USD)$0.0616
Per GiB-hour (USD)$0.0087

Full specification

Familya2
Archx86_64
Vcpu96
Memory Gi B680
Shared CpuNo
Image Space Gi B0
DescriptionAccelerator Optimized: 8 NVIDIA Tesla A100 GPUs, 96 vCPUs, 680GB RAM
Gpus8
Gpu Modelnvidia-tesla-a100
DeprecatedNo
Is CustomNo
Mem Per Vcpu Gi B7.083
Regionsasia-northeast1, asia-northeast3, asia-southeast1, europe-west4, me-west1, us-central1, us-east1, us-west1, us-west3, us-west4
Region Count10
Zone LimitedYes
Zone Noteasia-northeast1: 2/3 zones; asia-northeast3: 2/3 zones; europe-west4: 2/3 zones; me-west1: 2/3 zones; us-east1: 1/3 zones; us-west1: 1/3 zones; us-west3: 1/3 zones; us-west4: 1/3 zones

On-demand price shown for us-central1, USD, refreshed 2026-07-23T18:38:18+00:00. Prices and availability vary by region and OS — use the interactive table for the rest.