GCP Compute Instance

a3-highgpu-8g

The a3-highgpu-8g is a GPU-accelerated GCP Compute instance with 208 vCPUs and 1872 GB of memory, 8 H100 GPUs, powered by Intel. Ideal for deep learning training, high-performance computing, and GPU-intensive inference workloads. On-demand pricing starts at $87.8325/hr in us-central1. Spot pricing starts at $8.7045/hr in europe-north1.

Price History

Pricing by Region

Region On-Demand Spot Reserved 1Y
us-central1$87.832$52.555$63.792
us-east4$87.832$47.754$61.103
us-west1$87.832$52.555$63.792
asia-south1$91.346$47.200$63.408
europe-north1$96.712$8.704$67.133
europe-west1$96.712$45.405$67.133
us-east5$98.926$44.404$68.439
us-west4$98.926$42.389$68.821
europe-west2$100.129$9.914$69.505
asia-east1$101.701$54.755$70.596
europe-west9$101.886$29.008$70.725
europe-west3$103.642$62.185$72.145
asia-south2$105.496$29.240$73.230
australia-southeast1$109.791$26.118$76.593
europe-west4$111.817$55.399$79.656
asia-northeast3$112.794$37.447$78.297
asia-southeast1$113.555$64.179$78.824
australia-southeast2$114.182$10.362$79.260
asia-northeast1$125.789$20.635$87.534

Frequently Asked Questions

What is the a3-highgpu-8g?
The a3-highgpu-8g is a GPU-accelerated GCP Compute instance with 208 vCPUs and 1872 GB of memory, 8 H100 GPUs, powered by Intel. Ideal for deep learning training, high-performance computing, and GPU-intensive inference workloads. On-demand pricing starts at $87.8325/hr in us-central1. Spot pricing starts at $8.7045/hr in europe-north1.
What is the cheapest region for a3-highgpu-8g?
us-central1 at $87.832/hr on-demand.
How much does a3-highgpu-8g cost per month?
Starting at $64117.70/mo on-demand in us-central1, or $6354.27/mo with spot pricing.