g4dn.12xlarge
Amazon EC2 · G4dn family
The g4dn.12xlarge is a gpu & accelerated Amazon EC2 instance type with 48 vCPUs and 192 GiB of memory, 4 × NVIDIA T4, running on Intel Xeon Family with 50 Gigabit networking, starting at $3.912 per hour on demand.
- vCPUs
- 48
- memory
- 192 GiB
- gpu
- 4× T4
- on-demand
- $3.912
- spot
- $1.727
- 1-yr reserved
- $2.465
dedicated
4 GiB per vCPU
64 GiB total
$2,856/mo · us-east-1
−56% vs on-demand
−37% vs on-demand
g4dn.12xlarge price by region
23 regions · hourly list prices in USD · On-demand · Spot · 1-yr Reserved · 3-yr Reserved
| US East (N. Virginia)us-east-1cheapestreferenceNorth America | $3.912 | $1.727 | $2.465 | $1.69 |
| US East (Ohio)us-east-2cheapestNorth America | $3.912 | $2.039 | $2.465 | $1.69 |
| US West (Oregon)us-west-2cheapestNorth America | $3.912 | $1.463 | $2.465 | $1.69 |
| EU (Stockholm)eu-north-1Europe | $4.15 | $1.443 | $2.614 | $1.793 |
| Asia Pacific (Mumbai)ap-south-1Asia | $4.306 | $1.466 | $2.713 | $2.093 |
| Canada (Central)ca-central-1North America | $4.343 | $1.012 | $2.736 | $1.876 |
| EU (Ireland)eu-west-1Europe | $4.362 | $1.871 | $2.977 | $1.884 |
| EU (Paris)eu-west-3Europe | $4.574 | $0.9667 | $2.882 | $2.223 |
| EU (London)eu-west-2Europe | $4.577 | $1.583 | $3.124 | $2.224 |
| EU (Milan)eu-south-1Europe | $4.58 | $2.366 | $3.126 | $1.979 |
| US West (N. California)us-west-1North America | $4.694 | $1.297 | $3.303 | $2.281 |
| Middle East (Bahrain)me-south-1Middle East | $4.798 | $0.6993 | $3.275 | $2.073 |
| Asia Pacific (Seoul)ap-northeast-2Asia | $4.812 | $1.835 | $3.026 | $2.079 |
| EU (Frankfurt)eu-central-1Europe | $4.89 | $1.201 | $3.337 | $2.377 |
| AWS GovCloud (US-East)us-gov-east-1North America | $4.931 | – | $3.106 | $2.13 |
| AWS GovCloud (US)us-gov-west-1North America | $4.931 | – | $3.106 | $2.13 |
| Asia Pacific (Sydney)ap-southeast-2Oceania | $5.087 | $1.966 | $3.205 | $2.472 |
| Africa (Cape Town)af-south-1Africa | $5.19 | $1.427 | $3.542 | $2.242 |
| Asia Pacific (Osaka)ap-northeast-3Asia | $5.281 | $0.6505 | $3.604 | $2.681 |
| Asia Pacific (Tokyo)ap-northeast-1Asia | $5.281 | $1.502 | $3.604 | $2.681 |
| Asia Pacific (Singapore)ap-southeast-1Asia | $5.474 | $1.814 | $3.449 | $2.66 |
| Asia Pacific (Hong Kong)ap-east-1Asia | $6.024 | $2.065 | $3.795 | $2.602 |
| South America (Sao Paulo)sa-east-1South America | $6.649 | $1.03 | $4.538 | $3.375 |
Specifications
compute
- vCPUs
- 48
- Memory
- 192 GiB
- Memory per vCPU
- 4 GiB
- Processor
- Intel Xeon Family
- CPU vendor
- Intel
- Architecture
- x86-64
- Clock speed
- 2.5 GHz
- Burstable CPU
- No
storage & network
- Local storage
- 900 GB NVME
- Storage detail
- 900 GB NVMe SSD
- Network
- 50 Gigabit
- Peak bandwidth
- 50 Gbps
accelerators
- Model
- NVIDIA T4
- Count
- 4
- Memory (total)
- 64 GiB
- Architecture
- Turing
- Vendor
- NVIDIA
aws details
- API name
g4dn.12xlarge- Family
- G4dn
- Category
- GPU & accelerated
- Generation
- Current
- Regions
- 23 (Africa, Asia, Europe, Middle East, North America, Oceania, South America)
- Cheapest region
- US East (N. Virginia) · $3.912/hr
- Dedicated EBS throughput
- 7000 Mbps
- Enhanced networking
- Yes
- Normalization size factor
- 7.44
- Processor features
- Intel AVX; Intel AVX2; Intel AVX512; Intel Turbo
- Spot interruption rate (us-east-1)
- >20%
- Spot savings vs on-demand (us-east-1)
- 65%
g4dn.12xlarge equivalents on other clouds
closest match by vCPU, memory, CPU architecture, accelerator and class · Linux on-demand in each provider's reference region
| Provider | Instance | vCPU | Memory | Processor | On-demand | vs g4dn.12xlarge |
|---|---|---|---|---|---|---|
| AWS | g4dn.12xlarge | 48 | 192 GiB | 4 × NVIDIA T4 | $3.912 | – |
| Google Cloud | g2-standard-48NVIDIA L4 | 48 | 192 GiB | 4 × NVIDIA L4 | $4.002 | +2% |
| Akamai | RTX4000 Ada x4 MediumNVIDIA RTX 4000 Ada | 48 | 196 GiB | 4 × NVIDIA RTX 4000 Ada | $3.57 | -9% |
G4dn family sizes
| Size | vCPU | Memory | Accelerator | Local storage | On-demand | Monthly | Spot | 1-yr |
|---|---|---|---|---|---|---|---|---|
| g4dn.xlarge | 4 | 16 GiB | 1 × NVIDIA T4 | 125 GB NVME | $0.526 | $384 | $0.2204 | $0.331 |
| g4dn.2xlarge | 8 | 32 GiB | 1 × NVIDIA T4 | 225 GB NVME | $0.752 | $549 | $0.3063 | $0.474 |
| g4dn.4xlarge | 16 | 64 GiB | 1 × NVIDIA T4 | 225 GB NVME | $1.204 | $879 | $0.5264 | $0.759 |
| g4dn.8xlarge | 32 | 128 GiB | 1 × NVIDIA T4 | 900 GB NVME | $2.176 | $1,588 | $0.8729 | $1.371 |
| g4dn.12xlarge | 48 | 192 GiB | 4 × NVIDIA T4 | 900 GB NVME | $3.912 | $2,856 | $1.727 | $2.465 |
| g4dn.16xlarge | 64 | 256 GiB | 1 × NVIDIA T4 | 900 GB NVME | $4.352 | $3,177 | $1.564 | $2.742 |
| g4dn.metal | 96 | 384 GiB | 8 × NVIDIA T4 | 1.8 TB NVME (2 disks) | $7.824 | $5,712 | $3.523 | $4.929 |
g4dn.12xlarge FAQ
How much does g4dn.12xlarge cost?
g4dn.12xlarge costs $3.912 per hour on demand with Linux in US East (N. Virginia) (us-east-1), about $2,856 per month. Across 23 regions the Linux on-demand price ranges from $3.912 in US East (N. Virginia) to $6.649 in South America (Sao Paulo). Spot capacity costs $1.727 per hour (56% less), but can be reclaimed. A 1-yr Reserved commitment brings it to $2.465 per hour. With Windows it costs $6.12 per hour.
How many vCPUs and how much memory does g4dn.12xlarge have?
g4dn.12xlarge has 48 vCPUs and 192 GiB of memory (4 GiB per vCPU), running on Intel Xeon Family. It is a gpu & accelerated instance type in the G4dn family.
What GPU does g4dn.12xlarge have?
g4dn.12xlarge has 4 × NVIDIA T4 with 64 GiB of accelerator memory in total.
What are the equivalents of g4dn.12xlarge on other clouds?
The closest matches by vCPU, memory, CPU architecture, accelerator and instance class are on Google Cloud, g2-standard-48 (48 vCPU, 192 GiB, $4.002/hour in Iowa); on Akamai, RTX4000 Ada x4 Medium (48 vCPU, 196 GiB, $3.57/hour in Newark, NJ).