Standard_ND128isr_NDR_GB200_v6


Azure Virtual Machine: ND128isr_NDR_GB200_v6 / ND128isr NDR_GB200_v6 with 128 vCPUs and 864 GiB of memory. Available in 19 regions starting from $78,956.80 per month.
NameStandard_ND128isr_NDR_GB200_v6
DetailsStandard is recommended tier
N – GPU enabled
D – Training and inference scenarios for deep learning
128 – The number of vCPUs
NDR – Accelerator Type
i – Isolated size
s – Premium Storage capable
r – Remote direct memory access (RDMA) capable
v6 – version
vCPUs128
CPU ArchitectureArm64
Memory (GiB)864
Hyper-V GenerationsV2
ACUs0
GPUs4
GPU NameNvidia Blackwell GPU (192GB)
GPU RAM (GiB)192
GPU Total RAM (GiB)768
Max Network Interfaces8
RDMA Enabledyes
Accelerated Netyes
OS Disk Size1023 GiB
Res Disk Size1024 GiB
Max Disks16
Support Premium Diskyes
Combined IOPS is a sum of all attached disk's IOPs450000
Uncached Disk IOPS260000
Combined Write is a sum of all attached disk's write throughtput3815 MiB/Sec
Combined Read is a sum of all attached disk's read throughtput3815 MiB/Sec

Regional Prices

US Dollar ($)
Per Hour
Standard
Pay-as-you-go
Region Name
Region ID
Linux Price
Windows Price
East USeastus108.1600114.0480
East US 2eastus2108.1600114.0480
hidden-1hidden-1108.1600114.0480
West US 2westus2108.1600114.0480
hidden-2hidden-2124.4000130.2880
hidden-3hidden-3126.5400132.4280
hidden-4hidden-4129.8000135.6880
hidden-5hidden-5129.8000135.6880
hidden-6hidden-6133.0370138.9250
hidden-7hidden-7135.2000141.0880
hidden-10hidden-10140.6000146.4880
hidden-8hidden-8140.6000146.4880
hidden-9hidden-9140.6000146.4880
hidden-11hidden-11146.0000151.8880
hidden-12hidden-12151.4400157.3280
hidden-13hidden-13156.8400162.7280
hidden-14hidden-14163.3200169.2080
hidden-15hidden-15201.1600207.0480
hidden-16hidden-16216.3200222.2080
Activate subscription to see all regions and unlock all features

Best AI models you can run on this instance

Top open-weight chat models ranked by Intelligence Index that fit in Standard_ND128isr_NDR_GB200_v6's 768 GB of GPU memory, assuming the model is sharded across all 4 GPUs. VRAM is estimated from the parameter count at the selected quantization, so treat it as guidance rather than a guarantee.

Chat Models
FP16 (full precision)
Model
Creator
Intelligence
Total params
Active params
~VRAM
Qwen3.8 27BAlibaba logoAlibaba#3927B—65 GB
DeepSeek V4 FlashDeepSeek logoDeepSeek#41158B—379 GB
Hy 3Tencent logoTencent#77299B—717 GB
7 more models fit on Standard_ND128isr_NDR_GB200_v6
Unlock the full ranked list and FP8 / INT4 quantization with a CloudPrice subscription.

Similar Alternative VMs

Find Similar Instances In:
Microsoft Azure