ct3-hightpu-4t by Google Cloud Platform

ct3-hightpu-4t is a 96 vCPUs, 340 GB RAM, 2 vNUMA nodes server offered by Google Cloud Platform with 96 vCPUs, 340 GiB of memory and 0 GB of storage.
96 vCPU
340 GiB Memory
4 TPU
Spare SCore
-
(All-cores)
-
(Single-core)

Specifications

Server Metadata

Vendor ID
gcp
Server ID
700094
Name
ct3-hightpu-4t
Description
96 vCPUs, 340 GB RAM, 2 vNUMA nodes
Family
ct3
Hw Virt
-
Status
active
Observed At
2026-10-24T02:21:42.273491

Availability

REGION / IDSPOTONDEMAND

Processor

vCPUs
96
CPU Allocation
Dedicated
CPU Architecture
x86_64

System Resources and Accelerators

MEMORY
Memory Amount
340 GiB
ACCELERATORS
Accelerator Count
4
Accelerator Memory Min
32 GiB
Accelerator Memory Total
128 GiB
Accelerator Type
TPU
Accelerator Manufacturer
Google
Accelerator Family
TPU
Accelerator Model
v3
Accelerators
  • Google TPU v3 (Memory amount: 32768)
  • Google TPU v3 (Memory amount: 32768)
  • Google TPU v3 (Memory amount: 32768)
  • Google TPU v3 (Memory amount: 32768)
STORAGE
Storage Size
0 GB
Storages
    NETWORK
    Inbound Traffic
    0 GB/month
    Outbound Traffic
    0 GB/month
    IPv4
    0

    Server Description

    A specialized accelerator instance featuring ninety-six dedicated cores and four integrated tensor processing units for demanding machine learning workloads.

    GPU AcceleratedCompute Optimized

    Google Cloud Platform ct3-hightpu-4t is a specialized, GPU-accelerated server instance designed for high-throughput machine learning workloads. It features 96 dedicated x86_64 vCPUs and 340.0 GB of system memory distributed across 2 vNUMA nodes, yielding approximately 3.54 GB of RAM per core. The instance is equipped with four Google TPU v3 accelerators providing a total of 128 GB of VRAM, making it highly optimized for tensor processing, deep learning model training, and large-scale inference. No local storage is bundled with this instance, and it includes zero complimentary public IPv4 addresses. This hardware profile makes the ct3-hightpu-4t highly cost-efficient for specialized machine learning and data science workloads that require dedicated TPU acceleration rather than general-purpose compute resources.

    Economics

    Average Price per Region

    Prices per Zone

    Lowest Prices

    Alternatives

    Servers of the Same Family

    INSTANCEvCPUsMEMORYAccelerators

    Similar Servers

    INSTANCEVENDORvCPUsMEMORYAccelerators

    ct3-hightpu-4t FAQs