Build up your AI
with the most scalable platform

L 0
open source
Unlimited open source repositories
2 Private repositories
5 Concurrent instances
10GB Storage
+ $0.05 / GB / Month
Sign up
No credit card required
L 1
per seat
per month
Unlimited repositories
200 Concurrent instances
1TB Storage
+ $0.05 / GB / Month
Enterprise Edition
Interested in
using TensorPort
with enterprise level features on any infrastructure?

Scale your AI with the absolute lowest cost infrastructure

We provide the absolute lowest infrastructure cost so that you can evaluate all of your ideas.
If you find a cheaper source of compute let us know and we will match the price for you.
A1.C4.2xlarge 64 732 ? N/A $12.80 / hour
A1.P2.xlarge 32 488 ? 1x Tesla K80 $6.42 / hour
A1.P2.xlargeblessed 4 61 ? 8x Tesla K80 $0.80 / hour
Content for the second tab
Content for the third tab
Content for the fouth tab

Compare plans

Plan L0 L1 L2 L∞
Concurrent instances 5 100 500 Contact us
Included storage 10GB 1TB 10TB Contact us
Max storage 50GB 5TB
Private repositories 2
Public repositories
SLA 99% 99.9% 99.95% 99.95%
Support L0 L1 L2 L∞
Custom Apps

Which plan is right for me?

If you are working on small projects and do not require to run many distributed jobs at once, then use the L0. If you have many projects and run multiple distributed experiments, then use L1. If you need high availability and %99.9 uptime guarantee then use our enterprise plan L2 and L∞.

What is a blessed instance?

Blessed instances run with %99.9 uptime guarantee. This is in contrast to the regular instances that are significantly lower in cost but may be interrupted.

Can I get more concurrent instances?

We can provide up to 1000 concurrent instances. If 100 concurrent instances is not enough for you, please contact us for a customized plan or upgrade to enterprise plans.

Will I be locked in on TensorPort?

No. You can export all of your data and trained models at anytime at no cost. It is simple to stop or delete your account in your dashboard.