AWS g4dn.12xlarge Pricing and Specs
Amazon Web Services g4dn.12xlarge is a accelerated-computing instance with 96 vCPUs and 384.0 GB memory. This page provides comprehensive pricing comparison across all major regions, including on-demand, reserved, and spot pricing options.
Specifications
| Specification | Value |
|---|
| vCPU | 96 |
| Memory | 384.0 GB |
| Storage | NVMe SSD |
| Network Performance | Up to 25 Gbps |
| EBS Bandwidth | Up to 4,750 Mbps |
| Processor | Intel Xeon Platinum 8259CL |
Pricing by Region
On-Demand Pricing
| Region | Hourly | Monthly |
|---|
| US East (N. Virginia) | $12.62 | $9215.52 |
| US West (Oregon) | $13.26 | $9676.30 |
| US Central (Iowa) | $12.88 | $9399.83 |
| Canada Central | $13.63 | $9952.76 |
| EU West (Ireland) | $13.26 | $9676.30 |
Reserved Instance (1-Year No Upfront)
| Region | Hourly | Savings vs On-Demand |
|---|
| US East (N. Virginia) | $7.57 | 40% |
| US Central (Iowa) | $7.73 | 40% |
| EU West (Ireland) | $7.95 | 40% |
Spot Pricing
| Region | Hourly | Savings vs On-Demand |
|---|
| US East (N. Virginia) | $3.79 | 70% |
| US Central (Iowa) | $3.86 | 70% |
| EU West (Ireland) | $3.98 | 70% |
Use Cases
accelerated-computing instance are ideal for g4dn.12xlarge:
- ML model training and inference
- Graphics rendering
- Video transcoding
Equivalent Instances
| Provider | Instance | vCPU | Memory | Price Diff |
|---|
| AWS | g4dn.12xlarge | 96 | 384.0 | baseline |
| Azure | D2s_v5 | 2 | 8 | -14% |
| GCP | e2-standard-2 | 2 | 8 | -12% |
| Alibaba Cloud | ecs.g6-large | 2 | 8 | -14% |
Data Source
- API: Amazon Web Services Official Pricing API
- Fetched: 2026-07-23
- Methodology: Prices retrieved via official API and normalized to USD hourly/monthly rates
