| title | Virtual machine sizes overview |
|---|---|
| description | Lists the different instance sizes available for virtual machines in Azure. |
| author | mattmcinnes |
| ms.service | azure-virtual-machines |
| ms.subservice | sizes |
| ms.topic | concept-article |
| ms.workload | infrastructure-services |
| ms.date | 03/09/2026 |
| ms.author | mattmcinnes |
Applies to: ✔️ Linux VMs ✔️ Windows VMs ✔️ Flexible scale sets ✔️ Uniform scale sets
Azure Virtual Machine (VM) sizes are designed to provide a wide range of options for hosting your servers and their workloads in the cloud. Sizes are categorized into different families and types, each optimized for specific purposes. Users can choose the most suitable VM size based on their requirements, such as CPU, memory, storage, and network bandwidth.
This article describes what sizes are, gives an overview of the available sizes and shows different options for Azure virtual machine instances you can use to run your apps and workloads.
:::image type="content" source="./media/vm-series-video-play.png" alt-text="YouTube video for selecting the right size for your VM." link="https://www.youtube.com/watch?v=zKUBjtof6nU":::
Azure VM sizes follow specific naming conventions to denote varying features and specifications. Each character in the name represents different aspects of the VM. These include the VM family, number of vCPUs, and extra features like premium storage or included accelerators.
VM naming is further broken down into the 'Series' name and the 'Size' name. Size names include extra characters representing the number of vCPUs, type of storage, etc.
| Category | Description | Links |
|---|---|---|
| Type | Basic categorization by intended workload. | General purpose Compute optimized Memory optimized Storage optimized GPU accelerated FPGA accelerated |
| Series | Group of sizes with similar hardware and features. | Enter the 'Series' tab here. |
| Size | Specific VM configuration, including vCPUs, memory, and accelerators. | Enter the 'Size' tab here. |
Here's a breakdown of a 'General purpose, DCads_v5-series' size series.
:::image type="content" source="./media/size-series-breakdown.png" alt-text="Graphic showing a breakdown of the DCadsv5 VM size series with text describing each letter and section of the name.":::
1 Most families are represented using one letter, but others such as GPU sizes (ND-series, NV-series, etc.) use two.
2 Most subfamilies are represented with a single upper case letter, but others (such as Ebsv5-series) are still considered subfamilies of their parent family due to feature differences.
3 If no feature letter for a CPU is listed, the series uses Intel x86-64 CPUs. If the CPU is AMD, it's listed as a. If the CPU is ARM based (Microsoft Cobalt or Ampere Altra), it's listed as p.
4 There can be any number of extra features in a size name. There could be none (Dv5-series) or there could be three (Dplds_v6-series).
5 Version numbers only appear in the size name if there are multiple versions of the same series. If you're using the first version of a series (HB-series, B-series, etc.) it's often not included in the size name.
Note
Not all sizes will have subfamilies, support accelerators, or specify the CPU vendor. For more information on VM size naming conventions, see Azure VM sizes naming conventions.
Here's a breakdown of a 'Standard_DC8ads_v5' size in the 'DCadsv5-series'
:::image type="content" source="./media/size-instance-breakdown.png" alt-text="Graphic showing a breakdown of the DC8ads_v5 VM size with text describing each letter and section of the name.":::
1 Most families are represented using one letter, but others such as GPU sizes (ND-series, NV-series, etc.) use two.
2 Most subfamilies are represented with a single upper case letter, but others (such as Ebsv5-series) are still considered subfamilies of their parent family due to feature differences.
3 If no feature letter for a CPU is listed, the series uses Intel x86-64 CPUs. If the CPU is AMD, it will be listed as a. If the CPU is ARM based (Microsoft Cobalt or Ampere Altra), it will be listed as p.
4 There can be any number of extra features in a size name. There could be none (Dv5-series) or there could be three (Dplds_v6-series).
5 Spacers can show up multiple times in a size name such as in the ND_H100_v5-series. In this case they separate the GPU ID from the rest of the size name.
6 Version numbers only appear in the size name if there are multiple versions of the same series. If you're using the first version of a series (HB-series, B-series, etc.) it's often not included in the size name.
Note
Not all sizes will have subfamilies, support accelerators, or specify the CPU vendor. For more information on VM size naming conventions, see Azure VM sizes naming conventions.
This section contains a list of all current generation size series with tabs dedicated to each size family. Each group has a 'Series List' column with a linked list of all available size series, These links will bring you to the family page for that series, where you can find detailed information on each size in that series or go to the series' page for a list of sizes in that series.
To learn more about a size family, select the 'family' tab under each type section. There you can read a summary on the family, see the workloads it's recommended for, and view the full family page with specifications for all series in that family.
General purpose VM sizes provide balanced CPU-to-memory ratio. Ideal for testing and development, small to medium databases, and low to medium traffic web servers.
| Family | Workloads | Series List |
|---|---|---|
| A-family | Entry-level economical | Av2-series Previous-gen A-family series |
| B-family | Burstable | Bsv2-series Basv2-series Bpsv2-series Previous-gen B-family series |
| D-family | Enterprise-grade applications Relational databases In-memory caching Data analytics |
Dsv7 and Ddsv7 series Dlsv7 and Dldsv7 series Dasv7 and Dadsv7 series Dalsv7 and Daldsv7 series Dpsv6 and Dplsv6-series Dpdsv6 and Dpldsv6-series Dasv6 and Dadsv6-series Dalsv6 and Daldsv6-series Dsv6 and Ddsv6-series Dlsv6 and Dldsv6-series Dpsv5 and Dpdsv5-series Dplsv5 and Dpldsv5-series Dlsv5 and Dldsv5-series Dv5 and Dsv5-series Ddv5 and Ddsv5-series Dasv5 and Dadsv5-series Previous-gen D-family series |
| DC-family | D-family with confidential computing | DCasv5 and DCadsv5-series DCas_cc_v5 and DCads_cc_v5-series DCesv6-series DCedsv6-series DCasv6 and DCadsv6-series DCsv3 and DCdsv3-series Previous-gen DC-family |
[!INCLUDE a-family-summary]
View the full A family page [!INCLUDE a-family-workloads]
[!INCLUDE b-family-summary]
View the full B family page [!INCLUDE b-family-workloads]
[!INCLUDE d-family-summary]
View the full D family page [!INCLUDE d-family-workloads]
[!INCLUDE dc-family-summary]
View the full DC family page [!INCLUDE dc-family-workloads]
Compute optimized VM sizes have a high CPU-to-memory ratio. These sizes are good for medium traffic web servers, network appliances, batch processes, and application servers.
List of compute optimized VM size families:
| Family | Workloads | Series List |
|---|---|---|
| F-family | Medium traffic web servers Network appliances Batch processes Application servers |
Fasv7 and Fadsv7 series Famsv7 and Famdsv7 series Falsv7 and Faldsv7 series Fasv6, Falsv6, and Famsv6-series Fsv2-series Previous-gen F-family |
| FX-family | Electronic Design Automation (EDA) Large memory relational databases Medium to large caches In-memory analytics |
FX-series FXmsv2-series FXmdsv2-series |
To learn more about a specific size family or series, select the tab for that family and scroll to find your desired size series.
[!INCLUDE f-family-summary]
View the full F family page [!INCLUDE f-series-workloads]
[!INCLUDE fx-family-summary]
View the full FX family page [!INCLUDE fx-series-workloads]
Memory optimized VM sizes offer a high memory-to-CPU ratio that is great for relational database servers, medium to large caches, and in-memory analytics.
List of memory optimized VM sizes with links to each series' family page section:
| Family | Workloads | Series List |
|---|---|---|
| E-family | Relational databases Medium to large caches In-memory analytics |
Esv7 and Edsv7-series Easv7 and Eadsv7 series Esv6 and Edsv6-series Epsv6 and Epdsv6-series Easv6 and Eadsv6-series Ev5 and Esv5-series Edv5 and Edsv5-series Easv5 and Eadsv5-series Epsv5 and Epdsv5-series Previous-gen families |
| Eb-family | E-family with High remote storage performance | Ebsv6 and Ebdsv6-series Ebdsv5 and Ebsv5-series |
| EC-family | E-family with confidential computing | ECasv6 and ECadsv6-series ECasv5 and ECadsv5-series ECas_cc_v5 and ECads_cc_v5-series |
| M-family | Extremely large databases Large amounts of memory |
Mbsv3 and Mbdsv3-series Msv3 and Mdsv3 Medium Memory series Msv3 and Mdsv3 High Memory series Mdsv3 Very High Memory series Msv2 and Mdsv2-series Msv2 High Memory Series M-series |
| Other families | Older generation memory optimized sizes | Dnsv6 and Dndsv6-series Dnlsv6 and Dnldsv6-series Ensv6 and Endsv6-series Previous-gen families |
To learn more about a specific size family or series, select the tab for that family and scroll to find your desired size series.
[!INCLUDE e-family-summary]
View the full E family page [!INCLUDE e-series-workloads]
[!INCLUDE eb-family-summary]
View the full Eb family page [!INCLUDE eb-series-workloads]
[!INCLUDE ec-family-summary]
View the full EC family page [!INCLUDE ec-series-workloads]
[!INCLUDE m-family-summary]
View the full M family page [!INCLUDE m-series-workloads]
Storage optimized virtual machine (VM) sizes offer high disk throughput and IO, and are ideal for Big Data, SQL, NoSQL databases, data warehousing, and large transactional databases. Examples include Cassandra, MongoDB, Cloudera, and Redis.
List of storage optimized VM size families:
| Family | Workloads | Series List |
|---|---|---|
| L-family | High disk throughput and IO Big Data SQL and NoSQL databases Data warehousing Large transactional databases |
Lsv4-series Lasv4-series Laosv4-series Lsv3-series Lasv3-series Previous-gen L-family |
To learn more about a specific size family or series, select the tab for that family and scroll to find your desired size series.
[!INCLUDE l-family-summary]
View the full L family page [!INCLUDE l-series-workloads]
GPU optimized VM sizes are specialized virtual machines available with single, multiple, or fractional GPUs. These sizes are designed for compute-intensive, graphics-intensive, and visualization workloads.
List of GPU optimized VM size families:
| Family | Workloads | Series List |
|---|---|---|
| NC-family | Compute-intensive Graphics-intensive Visualization |
NC_RTXPRO6000BSE_v6-series NCads_H100_v5-series NCCads_H100_v5-series NCv3-series NCasT4_v3-series NC_A100_v4-series |
| ND-family | Large memory compute-intensive Large memory graphics-intensive Large memory visualization |
ND GB300-v6-series ND GB200-v6-series ND_MI300X_v5-series ND-H200-v5-series ND-H100-v5-series NDm_A100_v4-series ND_A100_v4-series |
| NG-family | Virtual Desktop (VDI) Cloud gaming |
NGads V620-series |
| NV-family | Virtual desktop (VDI) Single-precision compute Video encoding and rendering |
NV-series NVv3-series NVv4-series NVadsA10_v5-series NVads V710 v5-series Previous-gen NV-family |
To learn more about a specific size family or series, select the tab for that family and scroll to find your desired size series.
[!INCLUDE nc-family-summary]
View the full NC family page [!INCLUDE nc-series-workloads]
[!INCLUDE nd-family-summary]
View the full ND family page [!INCLUDE nd-series-workloads]
[!INCLUDE ng-family-summary]
View the full NG family page [!INCLUDE ng-series-workloads]
[!INCLUDE nv-family-summary]
View the full NV family page [!INCLUDE nv-series-workloads]
FPGA optimized VM sizes are specialized virtual machines available with single or multiple FPGAs. These sizes are designed for compute-intensive workloads. This article provides information about the number and type of FPGAs, vCPUs, data disks, and NICs. Storage throughput and network bandwidth are also included for each size in this grouping.
List of field programmable gate array accelerated VM size families:
| Family | Workloads | Series List |
|---|---|---|
| NP-family | Machine learning inference Video transcoding Database search and analytics |
NP-series |
To learn more about a specific size family or series, select the tab for that family and scroll to find your desired size series.
[!INCLUDE np-family-summary]
View the full NP family page [!INCLUDE np-series-workloads]
Azure High Performance Compute VMs are optimized for various HPC workloads such as computational fluid dynamics, finite element analysis, frontend and backend EDA, rendering, molecular dynamics, computational geoscience, weather simulation, and financial risk analysis.
List of high performance computing optimized VM size families:
| Family | Workloads | Series List |
|---|---|---|
| HB-family | High memory bandwidth Fluid Dynamics Weather modeling |
HBv2-series HBv3-series HBv4-series HBv5-series |
| HC-family | High density compute Finite element analysis Molecular dynamics Computational chemistry |
HC-series (retiring May 31, 2027) |
| HX-family | Large memory capacity Electronic Design Automation (EDA) |
HX-series |
To learn more about a specific size family or series, select the tab for that family and scroll to find your desired size series.
Note
HC-series (Standard_HC44rs, Standard_HC44-16rs, Standard_HC44-32rs) will be retired on May 31, 2027. After this date, remaining HC-series VMs will be deallocated, stop running, and will no longer have SLA or support. Sales of 1-year and 3-year Reserved Instances for HC-series ended on April 2, 2026. For new deployments, consider HBv5-series for higher performance and price-performance, or HX-series for high-memory HPC workloads.
[!INCLUDE hb-family-summary]
View the full 'HB' family page [!INCLUDE hb-series-workloads]
[!INCLUDE hc-family-summary]
View the full HC family page [!INCLUDE hc-series-workloads]
Note
Sales of 1-year and 3-year Reserved Instances for HC-series ended on April 2, 2026. HC-series is scheduled for retirement on May 31, 2027. Plan capacity accordingly.
[!INCLUDE hx-family-summary]
View the full HX family page [!INCLUDE hx-series-workloads]
The Azure CLI can be leveraged to identify which machine sizes are available for a particular region using the command:
az vm list-skus --location <region> --output table
- For information about pricing of the various sizes, see the pricing pages for Linux or Windows.
- Want to change the size of your VM? See Change the size of a VM.
- For availability of VM sizes in Azure regions, see Products available by region.
- To see general limits on Azure VMs, see Azure subscription and service limits, quotas, and constraints.
- For more information on how Azure names its VMs, see Azure virtual machine sizes naming conventions.
For information on using the REST API to query for VM sizes, see the following:
- List available virtual machine sizes for resizing
- List available virtual machine sizes for a subscription
- List available virtual machine sizes in an availability set
Learn more about compute performance for Linux VMs using the CoreMark benchmark scores.
Learn more about compute performance for Windows VMs using the SPECInt benchmark scores.
[!INCLUDE sizes-footer]