Gradient Subscriptions
An overview of Gradient Subscription pricing is available at https://gradient.paperspace.com/pricing.
Note: Upgraded Gradient Subscriptions provide access to Subscription benefits for each calendar month. If you wish to discontinue your subscription, you may cancel at any time but must downgrade to Free before Midnight (UTC) on the 1st to avoid the monthly subscription charge.
Private Workspace Options
G0 - Free (Beginner)
Each new Paperspace Account starts with a Free Gradient Subscription, or "G0". This is intended for those new to Machine Learning and/or those seeking to better understand how Paperspace Gradient works, without having to spend any money.
Notable limitations are:
- Auto-Shutdown is limited to 6 hours and is not configurable
- Only 1 Notebook and 1 Job can be run simultanelously
- 5GB Persistent Storage for Private workspaces
- Access only to Free and Low Instance types
G1 - $8/mo (Developer) and G2 - $24/mo (Professional)
These subscription types are designed to offer higher Service Limits and greater flexibility in your work. They offer far more Persistent Storage, configurable auto-shutdown, multiple concurrently-running Jobs and Notebooks, as well as access to higher-tier Instance Types. Visit our Gradient Pricing Page for full details.
Important: Gradient Community Notebooks / Free Instances have set limits regardless of your Gradient subscription level. Visit this doc for more information.
Team Workspace Options
T0 - Free (Small Teams)
Very similar to G0, the Private workspace equivalent, this free Gradient Subscription tier has limits. It is intended for small teams who would like to get a feel for Paperspace Gradient before upgrading.
T1 - $12/mo per user (Mid-size Teams) and T2 - $49/mo per user (Large Teams)
These paid subscription types are also similar to their Private workspace counterparts, offering access to higher instance types, greater storage, and auto-shutdown intervals.
Important: Team pricing is the monthly fee multiplied by however many Members exist on the team.
Notebook and Job Utilization
Notebooks and Jobs offer the same access to high-powered GPUs without maintaining and paying for a virtual machine and its storage. Usage is prorated down to the millisecond. You're billed for the time your Notebook or Job is using resources, according to the hourly rate of the machine you choose:
CPU Instances
A range of affordable and high-end CPU-only instances.
Name | Description | GPU | Memory | vCPUs | Price |
Low-costPreemptible price
|
---|---|---|---|---|---|---|
C2 | Low-cost instance with 1 vCPU, 1GB RAM | - | 1 GB | 1 vCPU | $0.009/hr | - |
C7 | Mid-range instance with 12 vCPU, 30GB RAM | - | 30 GB | 12 vCPU | $0.30/hr | - |
G1 | Low-cost instance with 1 vCPU, 1.7GB RAM | - | 1.7 GB | 1 vCPU | $0.0546/hr | $0.0418 |
G6 | Mid-rang instance with 6 vCPU, 24GB RAM | - | 24 GB | 6 vCPUs | $0.33/hr | $0.12 |
G12 | High-end instance with 12 vCPU, 64GB RAM | - | 64 GB | 12 vCPUs | $0.70/hr | $0.23 |
GPU Instances
A range of dedicated GPU instances with single and multi-GPU configurations.
Name | Description | GPU | Memory | vCPUs | Price |
Low-costPreemptible price
|
---|---|---|---|---|---|---|
GPU+ | Low-cost instance with 8GB GPU, 8 vCPU, 30GB RAM | 8 GB | 30 GB | 8 vCPU | $0.45/hr | - |
P4000 | Low-cost instance with 8GB GPU, 8 vCPU, 30GB RAM | 8 GB | 30 GB | 8 vCPU | $0.51/hr | - |
P5000 | Mid-range instance with 16GB Pascal GPU, 8 vCPU, 30GB RAM | 16 GB | 30 GB | 8 vCPU | $0.78/hr | - |
P6000 | High-performance instance with 24GB Pascal GPU, 8 vCPU, 30GB RAM | 24 GB | 30 GB | 8 vCPUs | $1.10/hr | - |
V100 | Ultra high-performance instance with 16GB Volta GPU, 8 vCPU, 30GB RAM | 16 GB | 30 GB | 8 vCPUs | $2.30/hr | - |
K80 | Low-cost instance with 12GB GPU, 2 vCPU, 12GB RAM | 12 GB | 12 GB | 2 vCPUs | $0.59/hr | $0.25/hr |
P100 | High-end instance with 16GB Pascal GPU, 4 vCPU, 24GB RAM | 16 GB | 24 GB | 4 vCPUs | $1.72/hr | $0.59/hr |
GV100 | Ultra high-performance instance with 16GB Volta GPU, 8 vCPU, 30GB RAM | 16 GB | 30 GB | 8 vCPUs | $2.89/hr | $1.15/hr |
GV100x8 | *Multi-GPU* instance with 128GB Volta GPU, 20 vCPU, 130GB RAM | 128 GB | 130 GB | 20 vCPUs | $20.99/hr | $8.43/hr |
TPU Instances
The most advanced AI accelerator chip available.
Name | Description | GPU | Memory | vCPUs | Price |
---|---|---|---|---|---|
TPUv2 | Low-cost instance with 8GB GPU, 8 vCPU, 30GB RAM | 16 GB | - | - | $8.42/hr |
*This is an additional GPU type for Gradient only.