Pricing & Compute

See the full list of compute options and pricing

On VESSL AI, you will be charged credits for each hour of computing. One credit is equivalent to $1.00. Core plan users are given 100 credits every month. They can buy additional credits under Organization Billing & Plan. Please refer to our Pricing page for more details or contact our sales team at [e-mail] if you need any help.

Resource spec preset

Resource spec preset refers to pre-defined resources on the VESSL-managed cluster. You can combine a cluster and preset in YAML to define resources. Below is a list resource spec presets of each VESSL-managed cluster, including the corresponding credits per hour.

VESSL-managed GCP cluster (Oregon)

Cluster name: vessl-gcp-oregon

Preset	GPUs	GPU Memory	vCPUs	Memory	Spot	Credits(/h)
cpu-medium	N/A	N/A	3	10 GB	False	0.134
cpu-medium-spot	N/A	N/A	3	10 GB	True	0.02
gpu-l4-small	1	24 GB	8	27 GB	False	1.177
gpu-l4-small-spot	1	24 GB	8	27 GB	True	0.7
gpu-l4-medium	2	48 GB	22	84 GB	False	2.354
gpu-l4-medium-spot	2	48 GB	22	84 GB	True	1.4
gpu-l4-large	4	96 GB	44	168 GB	False	4.708
gpu-l4-large-spot	4	96 GB	44	168 GB	True	2.8
gpu-v100-small	1	24 GB	8	27 GB	False	3.037
gpu-v100-small-spot	1	24 GB	8	27 GB	True	1.8

Example

GCP cluster in Oregon region with one L4 GPU spot instance

resources:
  cluster: vessl-gcp-oregon
  preset: gpu-l4-small-spot

VESSL-managed AWS cluster (Oregon)

Cluster name: vessl-aws-oregon

Preset	GPUs	GPU Memory	vCPUs	Memory	Spot	Credits(/h)
cpu-small	N/A	N/A	0.25	1 GB	False	0.0338
cpu-medium	N/A	N/A	2	6 GB	False	0.096
cpu-medium-spot	N/A	N/A	2	6 GB	True	0.0338
cpu-large	N/A	N/A	4	13 GB	False	0.192
gpu-t4-small	1	16 GB	4	13 GB	False	0.52
gpu-t4-small-spot	1	16 GB	4	13 GB	True	0.1578
gpu-t4-large	4	64 GB	48	163 GB	False	3.912
gpu-t4-large-spot	4	64 GB	48	163 GB	True	1.2719
gpu-a10g-small	1	24 GB	8	26 GB	False	1.212
gpu-a10g-small-spot	1	24 GB	8	26 GB	True	4.622
gpu-a10g-large	4	96 GB	96	362 GB	False	8.144
gpu-a10g-large-spot	4	96 GB	96	362 GB	True	5.7539
gpu-a10g-xlarge	8	192 GB	192	752 GB	False	16.288
gpu-a10g-xlarge-spot	8	192 GB	192	752 GB	True	11.5309
gpu-v100-small	1	16 GB	8	52 GB	False	3.06
gpu-v100-small-spot	1	16 GB	8	52 GB	True	0.918

Example

AWS cluster in Oregon region with one A10G GPU spot instance

resources:
  cluster: vessl-aws-oregon
  preset: gpu-a10g-small-spot

VESSL-managed AWS cluster (Seoul)

Cluster name: vessl-aws-seoul

Preset	GPUs	GPU Memory	vCPUs	Memory	Spot	Credits(/h)
cpu-small	N/A	N/A	0.25	1 GB	False	0.0315
cpu-medium	N/A	N/A	2	6 GB	False	0.0118
cpu-medium-spot	N/A	N/A	2	6 GB	True	0.0315
cpu-large	N/A	N/A	4	13 GB	False	0.236
cpu-large-spot	N/A	N/A	4	13 GB	True	0.0629
gpu-t4-small	1	16 GB	4	13 GB	False	0.647
gpu-t4-small-spot	1	16 GB	4	13 GB	True	0.1941
gpu-t4-large	4	64 GB	48	163 GB	False	4.812
gpu-t4-large-spot	4	64 GB	48	163 GB	True	1.4436
gpu-a10g-small	1	24 GB	8	26 GB	False	1.4903
gpu-a10g-small-spot	1	24 GB	8	26 GB	True	0.4471
gpu-a10g-large	4	96 GB	96	362 GB	False	10.014
gpu-a10g-large-spot	4	96 GB	96	362 GB	True	3.0042
gpu-a10g-xlarge	8	192 GB	192	752 GB	False	20.0281
gpu-a10g-xlarge-spot	8	192 GB	192	752 GB	True	6.3455
gpu-v100-small	1	16 GB	8	52 GB	False	4.234
gpu-v100-small-spot	1	16 GB	8	52 GB	True	1.2702
gpu-v100-large	4	64 GB	32	232 GB	False	16.936
gpu-v100-large-spot	4	64 GB	32	232 GB	True	5.0808
gpu-v100-xlarge	8	128 GB	64	460 GB	False	33.872
gpu-v100-xlarge-spot	8	128 GB	64	460 GB	True	10.1616

Example

AWS cluster in Korea region with one A10G GPU spot instance

resources:
  cluster: vessl-aws-seoul
  preset: gpu-a10g-small-spot

VESSL-manged on-premise cluster

Cluster name: vessl-kr-a100-80g-sxm

Preset	GPUs	GPU Memory	vCPUs	Memory	Spot	Credits(/h)
cpu-medium	N/A	N/A	2	16 GB	False	0.2
cpu-large	N/A	N/A	4	32 GB	False	0.4
gpu-a100-80g-small	1	80 GB	7	192 GB	False	1.8
gpu-a100-80g-medium	2	160 GB	14	384 GB	False	3.6
gpu-a100-80g-large	4	320 GB	28	768 GB	False	7.2
gpu-a100-80g-xlarge	8	640 GB	56	1536 GB	False	14.4

Example

VESSL-managed on-premise cluster with one A100 GPU instance

resources:
  cluster: vessl-kr-a100-80g-sxm
  preset: gpu-a100-80g-small

FAQ

How does VESSL calculate run usage?

We calculate the total workload hours on VESSL Run, Service, and Pipeline.

Do credits expire?

Yes, credits do expire. All credits you purchase have an expiration of one year. For the Enterprise plan, you can buy bulk credits with an extended expiration period.

Can I buy additional credits?

Yes, please visit Organization > Settings > Billing And Plan page, or reach out to our sales team at [e-mail].

Do you offer custom invoicing?

Yes, we offer custom invoicing for Enterprise customers. Contact our sales team([e-mail]) to learn more.