VM Types

VM types determine the hardware characteristics of your virtual machines (VMs).

You must specify a VM type when you create a VM, but you can also modify this type later. For more information, see Creating VMs and Modifying a VM Attribute.

TINA Type

The TINA type enables you to fully customize the compute values of your VM.

When you specify the TINA type, you must follow the tinavW.cXrYpZ or tinavW.cXrY naming pattern, replacing vW, X, Y, and Z with the values that you want according to the following limits:

Naming Pattern CPU Generation (vW) vCPUs (X) Memory (in GiB) (Y) Performance Flag (Z) Max. Private IPs per NIC

tinavW.cXrYpZ or tinavW.cXrY

Between v2 and v7

Between 1 and 78

Between 1 and 1039

Between 1 and 3

10

Predefined Types

Unlike the TINA type, predefined types are not customizable. They have specific names and specific compute values, and may also have specific physical characteristics in terms of CPU types of GPUs.

eu-west-2

Name CPU Generation vCPUs Memory (in GiB) Performance Flag Max. Private IPs per NIC Other Aspects

440fxlegacy1.large

v6

2

4

1

6

Intel 440FX chipset

c5.18xlarge

v5

72

147

1

30

c5.2xlarge

v5

8

16

1

15

c5.4xlarge

v5

16

33

1

30

c5.8xlarge

v5

36

74

1

30

c5.9xlarge

v5

36

72

1

30

c5.large

v5

2

4

1

10

c5.xlarge

v5

4

8

1

15

c7.2xlarge

v7

8

15

1

15

c7.4xlarge

v7

16

31

1

30

c7.8xlarge

v7

36

61

1

30

c7.large

v7

2

4

1

10

c7.xlarge

v7

4

8

1

15

g7d.xlarge

v7

8

32

1

30

1 GPU (nvidia-l40)

inference7-h100.2xlarge

v7

64

512

1

6

4 GPUs (nvidia-h100)

inference7-h100.large

v7

32

256

1

6

2 GPUs (nvidia-h100)

inference7-h100.medium

v7

16

128

1

6

1 GPU (nvidia-h100)

inference7-h100.xlarge

v7

64

450

1

6

2 GPUs (nvidia-h100)

inference7-h200.4xlarge

v7

128

512

1

6

8 GPUs (nvidia-h200)

inference7-h200.4xlargeA

v104

126

1024

1

6

8 GPUs (nvidia-h200)

inference7-h200.8xlarge

v7

128

2800

1

6

8 GPUs (nvidia-h200)

inference7-l40.large

v7

32

256

1

6

2 GPUs (nvidia-l40)

inference7-l40.medium

v7

16

128

1

6

1 GPU (nvidia-l40)

l40.2xlarge

v7

8

61

1

30

1 GPU (nvidia-l40)

m5.12xlarge

v5

48

197

2

30

m5.24xlarge

v5

96

384

2

30

m5.2xlarge

v5

8

33

2

15

m5.4xlarge

v5

16

66

2

30

m5.large

v5

2

8

2

10

m5.xlarge

v5

4

16

2

15

m7.10xlarge

v7

40

164

2

30

m7.2xlarge

v7

8

33

2

15

m7.4xlarge

v7

16

66

2

30

m7.large

v7

2

8

2

10

m7.xlarge

v7

4

16

2

15

p100.2xlarge

v5

8

61

1

15

1 GPU (nvidia-p100)

p100.4xlarge

v5

16

122

1

15

1 GPU (nvidia-p100)

p100.8xlarge

v5

32

244

1

15

2 GPUs (nvidia-p100)

p3.2xlarge

v5

8

61

1

15

1 GPU (nvidia-p3)

p6.8xlarge

v5

32

244

1

15

1 GPU (nvidia-p6)

r5.16xlarge

v5

64

500

2

30

r5.2xlarge

v5

8

62

2

15

r5.4xlarge

v5

16

125

2

30

r5.8xlarge

v5

32

250

2

30

r5.large

v5

2

16

2

10

r5.xlarge

v5

4

31

2

15

r7.16xlarge

v7

64

500

2

30

r7.2xlarge

v7

8

62

2

15

r7.4xlarge

v7

16

125

2

30

r7.8xlarge

v7

32

250

2

30

r7.large

v7

2

16

2

10

r7.xlarge

v7

4

31

2

15

us-east-2

Name CPU Generation vCPUs Memory (in GiB) Performance Flag Max. Private IPs per NIC Other Aspects

440fxlegacy1.large

v6

2

4

1

6

Intel 440FX chipset

c5.18xlarge

v5

72

147

1

30

c5.2xlarge

v5

8

16

1

15

c5.4xlarge

v5

16

33

1

30

c5.8xlarge

v5

36

74

1

30

c5.9xlarge

v5

36

72

1

30

c5.large

v5

2

4

1

10

c5.xlarge

v5

4

8

1

15

c7.2xlarge

v7

8

15

1

15

c7.4xlarge

v7

16

31

1

30

c7.8xlarge

v7

36

61

1

30

c7.large

v7

2

4

1

10

c7.xlarge

v7

4

8

1

15

g7d.xlarge

v7

8

32

1

30

1 GPU (nvidia-l40)

inference7-h100.2xlarge

v7

32

256

1

6

2 GPUs (nvidia-h100)

inference7-h100.large

v7

64

512

1

6

4 GPUs (nvidia-h100)

inference7-h100.medium

v7

16

128

1

6

1 GPU (nvidia-h100)

inference7-h200.4xlargeA

v104

126

1024

1

6

8 GPUs (nvidia-h200)

inference7-l40.large

v7

32

256

1

6

2 GPUs (nvidia-l40)

inference7-l40.medium

v7

16

128

1

6

1 GPU (nvidia-l40)

m5.12xlarge

v5

48

197

2

30

m5.24xlarge

v5

96

384

2

30

m5.2xlarge

v5

8

33

2

15

m5.4xlarge

v5

16

66

2

30

m5.large

v5

2

8

2

10

m5.xlarge

v5

4

16

2

15

m7.10xlarge

v7

40

164

2

30

m7.2xlarge

v7

8

33

2

15

m7.4xlarge

v7

16

66

2

30

m7.large

v7

2

8

2

10

m7.xlarge

v7

4

16

2

15

p100.2xlarge

v5

8

61

1

15

1 GPU (nvidia-p100)

p100.4xlarge

v5

16

122

1

15

1 GPU (nvidia-p100)

p100.8xlarge

v5

32

244

1

15

2 GPUs (nvidia-p100)

p3.2xlarge

v5

8

61

1

15

1 GPU (nvidia-p3)

p6.8xlarge

v5

32

244

1

15

1 GPU (nvidia-p6)

r5.16xlarge

v5

64

500

2

30

r5.2xlarge

v5

8

62

2

15

r5.4xlarge

v5

16

125

2

30

r5.8xlarge

v5

32

250

2

30

r5.large

v5

2

16

2

10

r5.xlarge

v5

4

31

2

15

r7.16xlarge

v7

64

500

2

30

r7.2xlarge

v7

8

62

2

15

r7.4xlarge

v7

16

125

2

30

r7.8xlarge

v7

32

250

2

30

r7.large

v7

2

16

2

10

r7.xlarge

v7

4

31

2

15

us-west-1

Name CPU Generation vCPUs Memory (in GiB) Performance Flag Max. Private IPs per NIC Other Aspects

440fxlegacy1.large

v6

2

4

1

6

Intel 440FX chipset

c5.18xlarge

v5

72

147

1

30

c5.2xlarge

v5

8

16

1

15

c5.4xlarge

v5

16

33

1

30

c5.8xlarge

v5

36

74

1

30

c5.9xlarge

v5

36

72

1

30

c5.large

v5

2

4

1

10

c5.xlarge

v5

4

8

1

15

c7.2xlarge

v7

8

15

1

15

c7.4xlarge

v7

16

31

1

30

c7.8xlarge

v7

36

61

1

30

c7.large

v7

2

4

1

10

c7.xlarge

v7

4

8

1

15

g7d.xlarge

v7

8

32

1

30

1 GPU (nvidia-l40)

inference7-h100.2xlarge

v7

32

256

1

6

2 GPUs (nvidia-h100)

inference7-h100.large

v7

64

512

1

6

4 GPUs (nvidia-h100)

inference7-h100.medium

v7

16

128

1

6

1 GPU (nvidia-h100)

inference7-h200.4xlargeA

v104

126

1024

1

6

8 GPUs (nvidia-h200)

inference7-l40.large

v7

32

256

1

6

2 GPUs (nvidia-l40)

inference7-l40.medium

v7

16

128

1

6

1 GPU (nvidia-l40)

m5.12xlarge

v5

48

197

2

30

m5.24xlarge

v5

96

384

2

30

m5.2xlarge

v5

8

33

2

15

m5.4xlarge

v5

16

66

2

30

m5.large

v5

2

8

2

10

m5.xlarge

v5

4

16

2

15

m7.10xlarge

v7

40

164

2

30

m7.2xlarge

v7

8

33

2

15

m7.4xlarge

v7

16

66

2

30

m7.large

v7

2

8

2

10

m7.xlarge

v7

4

16

2

15

p100.2xlarge

v5

8

61

1

15

1 GPU (nvidia-p100)

p100.4xlarge

v5

16

122

1

15

1 GPU (nvidia-p100)

p100.8xlarge

v5

32

244

1

15

2 GPUs (nvidia-p100)

p3.2xlarge

v5

8

61

1

15

1 GPU (nvidia-p3)

p6.8xlarge

v5

32

244

1

15

1 GPU (nvidia-p6)

r5.16xlarge

v5

64

500

2

30

r5.2xlarge

v5

8

62

2

15

r5.4xlarge

v5

16

125

2

30

r5.8xlarge

v5

32

250

2

30

r5.large

v5

2

16

2

10

r5.xlarge

v5

4

31

2

15

r7.16xlarge

v7

64

500

2

30

r7.2xlarge

v7

8

62

2

15

r7.4xlarge

v7

16

125

2

30

r7.8xlarge

v7

32

250

2

30

r7.large

v7

2

16

2

10

r7.xlarge

v7

4

31

2

15

cloudgouv-eu-west-1

Name CPU Generation vCPUs Memory (in GiB) Performance Flag Max. Private IPs per NIC Other Aspects

440fxlegacy1.large

v6

2

4

1

6

Intel 440FX chipset

c5.18xlarge

v5

72

147

1

30

c5.2xlarge

v5

8

16

1

15

c5.4xlarge

v5

16

33

1

30

c5.8xlarge

v5

36

74

1

30

c5.9xlarge

v5

36

72

1

30

c5.large

v5

2

4

1

10

c5.xlarge

v5

4

8

1

15

c7.2xlarge

v7

8

15

1

15

c7.4xlarge

v7

16

31

1

30

c7.8xlarge

v7

36

61

1

30

c7.large

v7

2

4

1

10

c7.xlarge

v7

4

8

1

15

g7d.xlarge

v7

8

32

1

30

1 GPU (nvidia-l40)

inference7-h100.2xlarge

v7

96

512

1

6

2 GPUs (nvidia-h100)

inference7-h100.large

v7

32

256

1

6

2 GPUs (nvidia-h100)

inference7-h100.medium

v7

16

128

1

6

1 GPU (nvidia-h100)

inference7-h100.xlarge

v7

64

450

1

6

2 GPUs (nvidia-h100)

inference7-h200.4xlargeA

v104

126

1024

1

6

8 GPUs (nvidia-h200)

inference7-l40.large

v7

32

256

1

6

2 GPUs (nvidia-l40)

inference7-l40.medium

v7

16

128

1

6

1 GPU (nvidia-l40)

m5.12xlarge

v5

48

197

2

30

m5.24xlarge

v5

96

384

2

30

m5.2xlarge

v5

8

33

2

15

m5.4xlarge

v5

16

66

2

30

m5.large

v5

2

8

2

10

m5.xlarge

v5

4

16

2

15

m7.10xlarge

v7

40

164

2

30

m7.2xlarge

v7

8

33

2

15

m7.4xlarge

v7

16

66

2

30

m7.large

v7

2

8

2

10

m7.xlarge

v7

4

16

2

15

p100.2xlarge

v5

8

61

1

15

1 GPU (nvidia-p100)

p100.4xlarge

v5

16

122

1

15

1 GPU (nvidia-p100)

p100.8xlarge

v5

32

244

1

15

2 GPUs (nvidia-p100)

p3.2xlarge

v5

8

61

1

15

1 GPU (nvidia-p3)

p6.8xlarge

v5

32

244

1

15

1 GPU (nvidia-p6)

r5.16xlarge

v5

64

500

2

30

r5.2xlarge

v5

8

62

2

15

r5.4xlarge

v5

16

125

2

30

r5.8xlarge

v5

32

250

2

30

r5.large

v5

2

16

2

10

r5.xlarge

v5

4

31

2

15

r7.16xlarge

v7

64

500

2

30

r7.2xlarge

v7

8

62

2

15

r7.4xlarge

v7

16

125

2

30

r7.8xlarge

v7

32

250

2

30

r7.large

v7

2

16

2

10

r7.xlarge

v7

4

31

2

15

ap-northeast-1

Name CPU Generation vCPUs Memory (in GiB) Performance Flag Max. Private IPs per NIC Other Aspects

440fxlegacy1.large

v6

2

4

1

6

Intel 440FX chipset

c5.18xlarge

v5

72

147

1

30

c5.2xlarge

v5

8

16

1

15

c5.4xlarge

v5

16

33

1

30

c5.8xlarge

v5

36

74

1

30

c5.9xlarge

v5

36

72

1

30

c5.large

v5

2

4

1

10

c5.xlarge

v5

4

8

1

15

c7.2xlarge

v7

8

15

1

15

c7.4xlarge

v7

16

31

1

30

c7.8xlarge

v7

36

61

1

30

c7.large

v7

2

4

1

10

c7.xlarge

v7

4

8

1

15

g7d.xlarge

v7

8

32

1

30

1 GPU (nvidia-l40)

inference7-h100.2xlarge

v7

32

256

1

6

2 GPUs (nvidia-h100)

inference7-h100.large

v7

64

512

1

6

4 GPUs (nvidia-h100)

inference7-h100.medium

v7

16

128

1

6

1 GPU (nvidia-h100)

inference7-h200.4xlargeA

v104

126

1024

1

6

8 GPUs (nvidia-h200)

inference7-l40.large

v7

32

256

1

6

2 GPUs (nvidia-l40)

inference7-l40.medium

v7

16

128

1

6

1 GPU (nvidia-l40)

m5.12xlarge

v5

48

197

2

30

m5.24xlarge

v5

96

384

2

30

m5.2xlarge

v5

8

33

2

15

m5.4xlarge

v5

16

66

2

30

m5.large

v5

2

8

2

10

m5.xlarge

v5

4

16

2

15

m7.10xlarge

v7

40

164

2

30

m7.2xlarge

v7

8

33

2

15

m7.4xlarge

v7

16

66

2

30

m7.large

v7

2

8

2

10

m7.xlarge

v7

4

16

2

15

p100.2xlarge

v5

8

61

1

15

1 GPU (nvidia-p100)

p100.4xlarge

v5

16

122

1

15

1 GPU (nvidia-p100)

p100.8xlarge

v5

32

244

1

15

2 GPUs (nvidia-p100)

p3.2xlarge

v5

8

61

1

15

1 GPU (nvidia-p3)

p6.8xlarge

v5

32

244

1

15

1 GPU (nvidia-p6)

r5.16xlarge

v5

64

500

2

30

r5.2xlarge

v5

8

62

2

15

r5.4xlarge

v5

16

125

2

30

r5.8xlarge

v5

32

250

2

30

r5.large

v5

2

16

2

10

r5.xlarge

v5

4

31

2

15

r7.16xlarge

v7

64

500

2

30

r7.2xlarge

v7

8

62

2

15

r7.4xlarge

v7

16

125

2

30

r7.8xlarge

v7

32

250

2

30

r7.large

v7

2

16

2

10

r7.xlarge

v7

4

31

2

15

Notes

Availability

The tables above only list the VM types that are publicly available. If you run the ReadVmTypes API method, other VM types may appear, but are reserved for specific purposes and with no guaranteed availability.

CPU Generation

The CPU generation is a value that corresponds to a family of CPUs:

  • v5: Intel Skylake

  • v6: Intel Ice Lake

  • v7: Intel Sapphire Rapids

  • v104: AMD

The older generations (v2, v3, and v4) have reached their End of Support and Maintenance (EOSM) states. They may be usable, but their availability is not guaranteed. For more information, see End-of-Life Policy.

Performance Flag

The performance flag is a value that determines the overall performance of the vCPUs:

  • 1 (highest): Ideal for intensive compute applications that require stable performances during the VM lifecycle. Note that flexible GPUs can only be used with this highest value. For more information, see About Flexible GPUs.

  • 2 (high): Offers varying performances over time while staying adapted to most use cases. This is the default value if the value is not specified.

  • 3 (medium): Does not guarantee consistent performances, but is ideal for machines with low CPU demand.

Related Page

Corresponding API Methods