Amazon SageMaker Canvas pricing

Overview

With Amazon SageMaker Canvas, you can interactively prepare, explore, and analyze data, and access ready-to-use models including support for Amazon Bedrock foundation models (FMs) and publicly available FMs from SageMaker JumpStart or create custom models to extract information and generate predictions.

With SageMaker Canvas, you pay for what you use. There are three factors that determine your bill: charges for the time you are running the SageMaker Canvas workspace instance, usage of ready-to-use models, and resources used to create custom models and generate predictions. As soon as you launch the Canvas application, a workspace instance is started and you are billed for the duration while the instance is running. Common tasks when you are logged into SageMaker Canvas include ingesting, preparing, and exploring data, experimenting with ML models, and viewing prediction and explainability results. AWS AI services are used when you select a ready-to-use model and you are charged for your use of the particular AWS AI service. For publicly available SageMaker JumpStart FMs, you are charged for the usage of the Amazon SageMaker instance on which the FM is deployed. When you create a custom model and generate predictions, you are charged depending on the type of custom model. When you deploy a Canvas model for real time inference you are charged for the usage of the Amazon SageMaker instance on which the model is deployed. For more information, see the Amazon SageMaker Pricing for Hosting: Real-Time Inference.

Amazon SageMaker Canvas expands access to machine learning (ML) by providing business analysts with a visual interface that allows them to generate accurate ML predictions on their own—without requiring any ML experience or having to write a single line of code. With Amazon SageMaker Canvas, you can interactively prepare, explore, and analyze data, and access ready-to-use models including support for Amazon Bedrock foundation models (FMs) and public FMs from SageMaker JumpStart or create custom models to extract information and generate predictions.

With SageMaker Canvas, you pay for what you use. There are three factors that determine your bill: charges for the time you are logged into the SageMaker Canvas workspace instance, usage of ready-to-use models, and resources used to create custom models and generate predictions. Workspace instance (Session-Hrs) charges are based on the amount of time you are logged into SageMaker Canvas, when resources are dedicated for you. Common tasks when you are logged into SageMaker Canvas include ingesting, preparing, and exploring data, experimenting with ML models, and viewing prediction and explainability results. AWS AI services are used when you select a ready-to-use model and you are charged for your use of the particular AWS AI service. For publicly available SageMaker JumpStart FMs, you are charged for the usage of the Amazon SageMaker instance on which the FM is deployed. When you create a custom model and generate predictions, you are charged depending on the type of custom model.

Workspace instance (Session-Hrs)

A workspace instance is dedicated for your use when you are logged into SageMaker Canvas. You pay based on the number of hours for which SageMaker Canvas is used or logged into. The time starts when you launch the SageMaker Canvas application, and ends either when you log out from the SageMaker Canvas interface or when your administrator ends your SageMaker Canvas application from the AWS management console. Logging out of SageMaker Canvas stops Workspace instance charges.

Workspace instance (Session-Hrs) charges
$1.9/hour

Ready-to-use Models

SageMaker Canvas offers ready to use models from Amazon Bedrock, Amazon Rekognition, Amazon Comprehend, and Amazon Textract for NLP and CV tasks. When you utilize a ready-to-use model, you will be charged directly from the respective service and their pricing terms and conditions apply.

Generation, extraction, and summarization of content using foundation models (FMs) from Amazon Bedrock is charged based on the volume of input tokens and output tokens. For more information, see Amazon Bedrock pricing. SageMaker JumpStart FMs are deployed on SageMaker instances. You are charged for the duration of usage based on the instance type used. For more information, see the Amazon SageMaker Pricing for Hosting: Real-Time Inference. Note: SageMaker Canvas automatically shuts down SageMaker JumpStart FMs after two hours of inactivity.

Requests for object detection in images and text detection in images are charged based on the number of images in your dataset. For further pricing information, please see Amazon Rekognition pricing.

Requests for sentiment analysis, entity extraction, language detection, and personal information detection are measured in units of 100 characters and you are charged based on the number of units in your dataset. For further pricing information, please see Amazon Comprehend pricing.

Requests for expense analysis, document analysis, and identity document analysis are measured in units of 1,000 pages and you are charged based on the number of units in your dataset. For further pricing information, please see Amazon Textract pricing.
 

Custom Models

For custom models, you are charged for SageMaker resources that are used to train your model and generate predictions.

Training charges

For custom models, you are charged for the resources that are used to train your model.

Custom tabular models

SageMaker Canvas supports numeric prediction (regression), 2 category prediction (binary classification), 3+ category prediction (multi-class classification), and time-series forecasting for tabular models. Canvas uses SageMaker Training and Processing instances, optimized for both latency and availability, to build your model. You are charged based on the instance hours used for model training in Amazon SageMaker. Canvas accelerates model training by using multiple SageMaker instances simultaneously to run parallel training jobs with different configurations. Therefore, the billable time reflects the total collective usage of these instances, which will likely be higher than the real-world clock time observed during a model build.

Training time and associated charges to build a model in SageMaker Canvas can vary based on several factors including dataset size, data type, model type, training method, and the specific instance types used for the job. You will see SageMaker training and processing charges for various instance types, including ml.m5.12xlarge, ml.c5.18xlarge, and ml.m5.4xlarge, that Canvas automatically selects based on performance and availability. The pricing for these instances can be found at Amazon SageMaker  pricing.


The following table may be used to estimate training cost of a Standard build, based on dataset size and SageMaker instance hours used. Please note that these figures are approximate, and actual hours and costs may vary.

Data Size Estimate SageMaker Training and Processing Instance hours Estimate charge

<100 MB

1 - 3

$2 - $8

100 MB -1 GB 3 - 11

$4 - $40

100 MB -1 GB 7 - 34 $16 - $100

 

The cost of training a Quick build, typically requiring between 0.07 to 0.7 SageMaker instance hours, may range from $0.20 to $2.00.

Custom NLP and CV models

SageMaker Canvas supports 2 and 3+ category prediction (binary and multiclass text classification and image classification) for custom NLP and CV models. The training charge for custom NLP and CV models is based on the amount of time it takes to train the model. SageMaker training instances are used to render the model training service and you will be charged directly from SageMaker. Based on the instances used from SageMaker Canvas, the training price will range from $2.03 - $4.89 per hour of training time. For further pricing information, please see SageMaker Pricing.

The following table provides estimated custom CV model training charges based on image resolution of 640 x 480 pixels. The estimates use a SageMaker instance price of $4.89/h. 

Number of images Estimate charge
100 $1.62
250 $1.63
500 $1.65
1,000 $1.68
5,000 $1.97
10,000 $2.33
50,000 $5.19

 

The following table provides estimated custom NLP model training charges based on an average of 240 unicode characters per cell. The estimates use a SageMaker instance price of $4.89/h.

Number of cells Estimate charge
100 $3.01
500 $3.11
1,000 $3.24
5,000 $4.22
10,000 $9.98
50,000 $15.25

 

Note: Training times and charges are subject to variance based on a number of factors including image resolution for CV, number of characters per sequence for NLP, and number of categories.

Prediction charges

For custom CV, NLP, and time-series forecasting models, you are charged for the resources that are used to generate your predictions. For other custom tabular models, there are no additional charges for predictions.

Custom tabular models

There is no additional charge for making predictions with numeric prediction, 2 category prediction, and 3+ category prediction custom tabular models. For predictions with time-series forecasting, charges apply for Amazon SageMaker Asynchronous Inference, Amazon SageMaker Batch Transform, or both.

Time-series forecasting models

With time-series forecasting models, you can generate either single or batch predictions.

For single prediction, SageMaker Asynchronous Inference charges apply, with a minimum of 2 hours. Depending on your region, charges may range from $0.408 to $0.533 per hour. The charges stop automatically after two idle hours.

For batch prediction, SageMaker Batch Transform charges apply based on the amount of time it takes to generate your predictions. The table below estimates charges based on the number of time-series observed in your data.

Number of time-series Estimate charge
1,000 $0.60
10,000 $9
100,000 $25

For detailed SageMaker pricing, please see SageMaker Pricing.

Custom NLP and CV models

The prediction charge for custom NLP and CV models is based on the amount of time it takes to generate your predictions. SageMaker instances, priced at $0.408 per hour of prediction generation time, are used to render the model prediction and you will be charged directly from SageMaker. For further pricing information, please see SageMaker Pricing.

For example, the estimated charge to generate a prediction for 1,000 images of resolution 640 x 480 is $0.03. Similarly, the estimated charge to generate a prediction for 1,000 sequences of 520 unicode characters per sequence is $0.01.

AWS Free Tier summary

Amazon SageMaker Canvas provides a 2-month free tier. The free tier includes workspace instance (Session-Hrs) usage up to 160 hours/month for using the SageMaker Canvas application.

Ready-to-use NLP and CV models are rendered by Amazon Rekognition, Amazon Comprehend, and Amazon Textract. Each service offers a varying free tier duration and coverage. To learn more, please see the respective AWS service price pages: Amazon Rekognition, Amazon Comprehend, and Amazon Textract.

Pricing table

Session  
$1.9 per hour  

 

Model Training  
First 10 Million cells $30 per million cells
Next 90 Million cells $15 per million cells
Over 100 Million cells $7 per million cells

Pricing examples

Example 1:
Let’s say you have a team of 4 analysts who want to try SageMaker Canvas. Let’s say one of them builds a numeric prediction model to predict on time delivery of packages, using a 50 MB input dataset. SageMaker Canvas used 2.9 instances hours of the ml.m5.12xlarge type to train the model. Through this process, the team is logged into SageMaker Canvas for 10 hours per week per user. The time is spent exploring data, preparing datasets, and generating predictions, translating to 40 hours per month per user, or 160 hours total usage. The bill at the end of the month would be calculated as follows:

Workspace instance (Session-Hrs) charges under free tier up to 160 hours/month: $0.00
Model training charges:  $2.765/hour x 2.9 = $7.69
Total: $7.69

Example 2:
Let’s say that after you consumed the free tier, your team continues to use SageMaker Canvas. You build a numeric prediction model using input date set of 150 MB. SageMaker Canvas used 10 instance hours of the ml.c5.18xlarge instance types to train the model. Throughout this process the team is logged into SageMaker Canvas and spends 40 hours in SageMaker Canvas during one month to explore data, join datasets, and run predictions. The bill at the end of the month would be calculated as follows:

Workspace instance (Session-Hrs) charges: $1.9 x 40 = $76
Model training charges:  $3.672/hour x 11 = $36.72
Total: $112.72

Example 3:
Let’s say after you consumed the free tier, you build a custom CV classification model to detect manufacturing defects in images and you use a training dataset of 1,000 images. Your training time is approximately 21 minutes and the price point is $4.89 per hour. During the process, you spend 4 hours in SageMaker Canvas to label the images in the training dataset, view the explainability heat map, and understand the model accuracy. You then run predictions which take approximately 12 minutes at a price point of $0.408 per hour. The bill would be calculated as follows:

Workspace instance (Session-Hrs) charges: $1.9*4 = $7.60
Model training charges: $4.89/hour x 21 mins x 1/60 = $1.68
Predictions: $0.408/hour x 12 mins x 1/60 = $0.08
Total: $9.36

Example 4:
Let’s say after you consumed the free tier, you build a custom NLP model to understand user sentiment in reviews and you use a training dataset of 6,700 reviews at an average of 120 characters per review and you use your model to generate predictions on 1,000 reviews. Your training time is approximately 31 minutes and the price point is $3.825 per hour, and the time to generate a prediction is 4.1 minutes and the price point is $0.408 per hour. During the process, you spend 2 hours in SageMaker Canvas to label the reviews in the training dataset and view the prediction results. The bill would be calculated as follows:

Workspace instance (Session-Hrs) charges: $1.9*2 = $3.80
Model training charges: $3.825/hour x 31 mins x 1/60 = $1.98
Predictions: $0.408/hour x 4.1 mins x 1/60 = $0.03
Total: $5.81

Example 5:
Let’s say after you consumed the free tier, you want to extract information from 50 identity documents. During the process, you spend 1.5 hours in SageMaker Canvas to import your documents and view your results. The bill would be calculated as follows:

Workspace instance (Session-Hrs) charges: $1.9*1.5 = $2.85
Ready-to-use model charges (as per Amazon Textract pricing): The pricing per page in the US West (Oregon) Region for the first 100,000 pages is $0.025 per page. The charge would be $0.025 x 50 = $1.25
Total: $4.10 

Example 6:

Let’s say that after you consumed the free tier, you build a custom time-series forecasting model to predict product demand. You own a clothing company with 1,000 items sold in 50 stores worldwide and are forecasting product demand for the next 12 weeks. You used a 200 MB dataset that includes weekly sales from the past year and information on two additional attributes: price and marketing spend. SageMaker Canvas uses 3 SageMaker Training instance hours of ml.m5.12xlarge instance types to train the model. After the model is built, you spend 30 minutes performing 'what-if' analysis with single predictions, which use SageMaker Asynchronous Inference on an ml.c5.2xlarge instance that SageMaker Canvas automatically stops after two idle hours. Subsequently, you generate batch predictions of a 12-week forecast horizon, requiring 3 SageMaker Batch Transform hours on ml.m5.12xlarge instances. Throughout this process, your team is logged into SageMaker Canvas, spending 10 hours during the month to explore data, join datasets, and run predictions. The bill at the end of the month would be calculated as follows:

Workspace instance (Session-Hrs) charges: $1.90 x 10 = $19
Model training charges:  $2.765/hour x 3 hrs = $8.30
Single Predictions: $0.408 x (30 mins of usage + 2 hrs idle time) = $1.02
Batch Predictions: $2.765/hour x 3 hrs = $8.30
Total = $36.62

Additional pricing resources