The pricing is dependent on the Service Level. We aim to have a flexible scheme where customers always pay on-demand for the value they extract.
An easy way to get started and test-drive the platform.
A good way to do some work and get used to our tools.
The perfect way to collaborate and get real work done.
Need private clusters, managed cloud, and extended privacy?
Below is a quick comparison of our pricing for different Service Levels.
The compute price above refers to the Ordinary cost category for each service level. One can further control the price by varying the category type. When using submission queues with "Saving" cost category, for example, the relative unit price can be as low as 1/5th of the Ordinary.
As the type of hardware and scheduling policies vary for different submission queues, the pricing is also different. For, example, GPU-enabled nodes are available within a certain subset of queues and are generally priced higher.Detailed description of submission queues is available HERE.Clusters and associated hardware and pricing are explained in this SECTION.
The options explained above can be combined in order to achieve the least expensive pricing. For example, when "Enterprise" service level is used in combination with submission queues that belong to the "Saving" cost category, the resulting price per core hour can be as low as $0.024. When we take into account the performance per core benchmarks this presents a unique performance per price option.