Azure VM Size Finder

Tell it the vCPU, memory and workload you need and get a ranked, priced shortlist of Azure VM sizes, in your region or wherever it is cheapest.

29 sizes match

32+ GiB, AI / ML (NVIDIA GPU), Linux, priced in East US

Understood from your description: 32+ GiB AI / ML (NVIDIA GPU) (“LLM”) East US

Cheapest size that fits

Standard_NV6ads_A10_v5

6 vCPU 55 GiB 1/6 A10 180 GiB temp disk x64 GPU, NVadsA10v5 series
$331.42/month $0.4540/hour, Linux pay as you go, East US
$145.83/month with a 3-year reservation

Alternatives

Size vCPU Memory GPU Temp disk Per month 3-yr reserved Compare
Standard_NC8as_T4_v3 NCasT4v3 · +24 GiB 8 56 GiB 1× T4 352 GiB $548.96 $206.42 vs #1
Standard_NV12ads_A10_v5 NVadsA10v5 · +78 GiB 12 110 GiB 1/3 A10 360 GiB $662.84 $291.64 vs #1
Standard_NC24lds_xl_RTXPRO6000BSE_v6 NCldsxlRTXPRO6000BSEv6 · +40 GiB 24 72 GiB 1× RTX PRO 6000 256 GiB $824.90 $362.94 vs #1
Standard_NC16as_T4_v3 NCasT4v3 · +78 GiB 16 110 GiB 1× T4 352 GiB $878.92 $330.47 vs #1
Standard_NC36lds_xl_RTXPRO6000BSE_v6 NCldsxlRTXPRO6000BSEv6 · +40 GiB 36 72 GiB 1× RTX PRO 6000 256 GiB $907.39 $399.25 vs #1
Standard_NC36ds_xl_RTXPRO6000BSE_v6 NCdsxlRTXPRO6000BSEv6 · +100 GiB 36 132 GiB 1× RTX PRO 6000 256 GiB $1,052.66 $463.14 vs #1
Standard_NV18ads_A10_v5 NVadsA10v5 · +188 GiB 18 220 GiB 1/2 A10 720 GiB $1,168.00 $513.92 vs #1
Standard_NC72lds_xl_RTXPRO6000BSE_v6 NCldsxlRTXPRO6000BSEv6 · +100 GiB 72 132 GiB 1× RTX PRO 6000 512 GiB $1,781.20 $783.72 vs #1
Standard_NC72ds_xl_RTXPRO6000BSE_v6 NCdsxlRTXPRO6000BSEv6 · +232 GiB 72 264 GiB 1× RTX PRO 6000 512 GiB $2,065.90 $909.11 vs #1
Standard_NC6s_v3 NCsv3 · +80 GiB 6 112 GiB 1× V100 736 GiB $2,233.80 — vs #1
Standard_NV36ads_A10_v5 NVadsA10v5 · +408 GiB 36 440 GiB 1× A10 1,440 GiB $2,336.00 $1,027.83 vs #1
Standard_NC64as_T4_v3 NCasT4v3 · +408 GiB 64 440 GiB 4× T4 2,816 GiB $3,176.96 $1,194.53 vs #1
Standard_NV36adms_A10_v5 NVadmsA10v5 · +848 GiB 36 880 GiB 1× A10 2,880 GiB $3,299.60 $1,451.83 vs #1
Standard_NC144lds_xl_RTXPRO6000BSE_v6 NCldsxlRTXPRO6000BSEv6 · +232 GiB 144 264 GiB 1× RTX PRO 6000 1,024 GiB $4,015.00 $1,766.61 vs #1
Standard_NC12s_v3 NCsv3 · +192 GiB 12 224 GiB 2× V100 1,474 GiB $4,467.60 — vs #1
Standard_NC144ds_xl_RTXPRO6000BSE_v6 NCdsxlRTXPRO6000BSEv6 · +484 GiB 144 516 GiB 1× RTX PRO 6000 1,024 GiB $4,657.40 $2,049.25 vs #1
Standard_NV72ads_A10_v5 NVadsA10v5 · +848 GiB 72 880 GiB 2× A10 2,880 GiB $4,759.60 $2,094.22 vs #1
Standard_NC40ads_H100_v5 NCadsH100v5 · +288 GiB 40 320 GiB 1× H100 128 GiB $5,095.40 $2,802.47 vs #1
Standard_NC288lds_xl_RTXPRO6000BSE_v6 NCldsxlRTXPRO6000BSEv6 · +484 GiB 288 516 GiB 2× RTX PRO 6000 2,048 GiB $8,030.00 $3,533.19 vs #1
Showing 20 of 29. USD list prices, 730 hours a month, updated 2 Nov 2026. Show 50

How the finder works

Sizes and specs

Every VM size Azure lists with az vm list-skus in 24 regions, refreshed weekly: vCPU, memory, GPUs, temporary disk, Premium SSD and accelerated networking support.

Prices

USD pay-as-you-go list prices from the Azure Retail Prices API for Linux or Windows (license included). With no region selected, each size is priced in its cheapest region.

Ranking

Lowest price lists the cheapest sizes that meet every requirement. Closest fit puts sizes with the least spare vCPU and memory first, then price.

Left out by default

Burstable B-series and confidential sizes appear only when you pick that workload. Constrained-vCPU sizes such as E8-2s_v5, used for per-core licensing, are not listed.

Frequently asked questions

How do I choose the right Azure VM size?

Start from the minimum vCPU and memory the workload needs, then pick the workload type and region. You can also describe it in plain English, for example "16 cores, 128 GB, SQL Server, West Europe", and the finder fills in the requirements. It lists every size that meets all of them, cheapest first, priced in your region or in each size's cheapest region.

Which Azure VM sizes are best for AI and machine learning?

Choose the AI / ML workload: it lists only NVIDIA GPU sizes (T4, A10, A100, H100, H200, RTX PRO 6000 and others), where CUDA, PyTorch and vLLM run. Small sizes such as Standard_NV6ads_A10_v5 get a slice of one GPU (1/6 of an A10, 4 GiB of GPU memory), enough for light inference. AMD Radeon sizes (V710, V620) are built for virtual desktops, graphics and cloud gaming and do not run CUDA.

Why are B-series burstable sizes not shown?

Burstable sizes earn CPU credits while idle and slow down to a baseline when the credits run out, so they only appear when you choose the Burstable workload. They suit dev/test machines, small web sites and jump boxes, not steady production load.

Are these the prices I will pay?

They are Microsoft's public USD pay-as-you-go list prices from the Azure Retail Prices API, refreshed every Monday. Your bill can be lower with an enterprise agreement, reservations, savings plans or Azure Hybrid Benefit, and it also includes disks, bandwidth and other services that are not priced here.

Is the cheapest region always the best choice?

No. Latency to your users, data residency rules and where the rest of your resources run usually matter more than a few dollars a month. Pick your region to compare sizes like for like, and use the size pages to see the price in every region.