Individual

PaaS

$5/month + Pay As You Go

  • building machine learning models
  • infrastructure and environment management
  • reproducibility
  • up to 4 concurrent jobs
  • 5GB of storage
  • standard support
Start free trial

60 days, no credit card required
free $100 for infrastructure,

Team

PaaS

$99/month + Pay As You Go

  • all features available
    in the individual tier
  • team collaboration
  • up to 20 concurrent jobs/user
  • 100GB of storage
  • standard support
◄ plan details

team contact

Enterprise

PaaS or on-premises

  • all features available in the team tier
  • machine learning model deployment
  • support for multiple cloud vendors
    (GCP, AWS and Azure)
  • integration with your existing cloud accounts
  • premium support
◄ plan details

enterprise contact

Compute Resources

Neptune integrates with the most popular public clouds, allowing you to expand your compute resources with a single command. There are many types of instances you can use, but the following common rules apply:

  • you are billed per second
  • you pay only for the time your experiment is running and uploading the results to the storage
  • the minimum charge for any experiment running in a cloud is $0.01

Instances on Google Cloud Platform (GCP)

NamevCPUsMemoryDisk
(SSD)
GPU'sGPU
memory
Cost($)
per second
Cost($)
per hour
gcp-small16.5GB100GB--0.00002530.0910800
gcp-medium426GB100GB--0.00007870.2833200
gcp-large852GB100GB--0.00015100.5436000
gcp-gpu-medium426GB100GB112GB0.00021480.7732800
gcp-gpu-large852GB100GB112GB0.00028711.0335600

Storage

Neptune uses storage to keep track of your source code, machine learning models and the outputs your experiments generate. Additionally, using storage you can exchange data between your jobs and notebooks.

You are charged $0.01 for each GB used per day (calculated as a maximum amount of GB in the day), that exceeds the storage included for your pricing tier.