Pricing

Video classification

Cloud AutoML Video Intelligence Classification enables you to train custom video machine learning models to classify videos into a custom set of categories. Prices for AutoML Video are based on resource usage, for both training and classification prediction.

Prices are listed in US Dollars (USD). If you pay in a currency other than USD, the prices listed in your currency on Cloud Platform SKUs apply.

Video classification training costs

The cost for training a model is $2.94 per node hour. For each unit of time, training a model requires the use of 9 nodes in parallel, where each node is equivalent to a n1-standard-4 machine with an attached NVIDIA® Tesla® V100 GPU.

You receive 40 node hours of free training credit per account (lifetime). Once used, the free training credit doesn't renew.

You pay only for the computer hours used; if training fails for any reason other than a user-initiated cancellation, you are not billed for the time. You will be charged for training time if you cancel the operation.

Video classification prediction costs

The cost for prediction is $0.42 per node hour. For each unit of time, prediction requires the use of 1 node equivalent to a n1-standard-8 machine.

You receive 5 node hours of free prediction credit per account (lifetime). Once used, the free prediction credit doesn't renew.

Prediction on 1 hour of video can cost less than 1 node hour, depending on the model and the video being annotated.

Video object tracking

Cloud AutoML Video Intelligence Object Tracking enables you to train custom video machine learning models to localize and track objects in your videos. Prices for AutoML Video are based on resource usage, for both training and prediction.

Prices are listed in US Dollars (USD). If you pay in a currency other than USD, the prices listed in your currency on Cloud Platform SKUs apply.

Video object tracking training costs

The cost for training a model is $2.94 per node hour. For each unit of time, training a model requires the use of 12 nodes in parallel, where each node is equivalent to a n1-standard-4 machine with an attached NVIDIA® Tesla® V100 GPU.

You receive 40 node hours of free training credit per account (lifetime). Once used, the free training credit doesn't renew.

You pay only for the computer hours used; if training fails for any reason other than a user-initiated cancellation, you are not billed for the time. You will be charged for training time if you cancel the operation.

Video object tracking prediction costs

The cost for prediction is $0.42 per node hour. For each unit of time, prediction requires the use of 1 node equivalent to a n1-standard-8 machine.

You receive 5 node hours of free prediction credit per account (lifetime). Once used, the free prediction credit doesn't renew.

Prediction on 1 hour of video can cost less than 1 node hour, depending on the model and the video being annotated.

Google Cloud Platform costs

You may be charged for other Google Cloud Platform resources used in your project, such as Compute Engine instances, Cloud Storage etc. See the Google Cloud Platform Pricing Calculator to determine other costs based on current rates.

To view your current billing status in the Google Cloud Platform Console, including usage and current bill, see the billing page. To learn more about billing, read the Google Cloud Platform billing documentation.

Was this page helpful? Let us know how we did:

Send feedback about...

AutoML Video Intelligence