ml.g4dn.16xlarge

    $5.4400/hr

    ~$3971.20/month on-demand in N. Virginia (us-east-1)

    The ml.g4dn.16xlarge is a GPU Accelerated SageMaker ML instance with 64 vCPUs and 256 GiB of memory with 1 GPU, starting at $5.440 per hour.

    Last updated: Feb 11, 2026, 05:00 AM UTC

    All pricing cards update to show this payment type

    Pricing Options

    On-Demand

    Pay as you go - no commitment

    Hourly Rate
    $5.4400/hr
    ~$3971.20/mo

    Reserved 1yr

    One year commitment

    Not available

    Reserved 3yr

    Three year commitment

    Not available

    G4dn Family

    Switch between sizes in N. Virginia (us-east-1)

    SizevCPUMem
    ml.g4dn.12xlarge48192
    ml.g4dn.16xlarge64256
    ml.g4dn.2xlarge832
    ml.g4dn.4xlarge1664
    ml.g4dn.8xlarge32128
    ml.g4dn.xlarge416

    Instance Specifications

    Detailed hardware and capabilities for ml.g4dn.16xlarge

    Processor

    vCPUs64
    Memory256 GiB
    Instance FamilyG4DN
    Architecture64-bit
    Clock Speed2.5 GHz
    Current Gen

    GPU & Accelerator

    Has GPU
    GPU Count1
    GPU NameT4
    GPU ManufacturerNVIDIA
    GPU Memory16 GiB

    Networking

    Network Performance50 Gigabit
    Network Bandwidth50 Gbps
    Max Network Interfaces4
    IPv6 Support
    ENA Support

    EBS (Elastic Block Store)

    EBS Optimizeddefault
    Baseline IOPS40,000 IOPS
    Max IOPS40,000 IOPS
    Baseline Bandwidth9,500 Mbps
    Max Bandwidth9,500 Mbps

    Instance Storage

    Instance Storage
    Storage Size900 GB
    Storage Typessd
    NVMe

    Regional Pricing Comparison

    Compare ml.g4dn.16xlarge pricing across all regions - sorted by cost

    Current region (us-east-1)Other regions
    Click any bar to open that region in a new tab