Google Cloud Vertex AI Pricing
Google Cloud Vertex AI offers a scalable and flexible pricing structure designed to meet diverse AI needs. Pricing packages include the Imagen Model for Image Generation, Text, Chat, and Code Generation, and data-specific training and prediction services.
A free trial with $300 in credits and a freemium version are available. Key features include data management, model training and deployment, prediction and analysis, and integration with Google Cloud services.
Pricing varies by service type and configuration, with additional support options and customizable pricing to optimize costs.
Google Cloud Vertex AI Plans
AutoML models
- Training (Edge on-device model): $18.00 per node hour (classification)
- Deployment and online prediction: $1.375 per node hour (classification)
- Batch prediction: $2.222 per node hour (classification)
Vertex AI Forecast
- Prediction:
- $0.1 per 1K data points* (1M-50M points)
- $0.02 per 1K data points* (>50M points)
- Training: $21.25/hr in all regions
Custom-trained models
- Iowa (us-central1): starting from n1-standard-4 - $0.218499
- Oregon (us-west1): starting from n1-standard-4 - $0.218499
- Northern Virginia (us-east4): starting from n1-standard-4 - $0.246100
- South Carolina (us-east1): starting from n1-standard-4 - $0.218499
- Montréal (northamerica-northeast1): starting from n1-standard-4 - $0.240580
- Singapore (asia-southeast1): starting from n1-standard-4 - $0.269560
- Sydney (australia-southeast1): starting from n1-standard-4 - $0.310155
- Taiwan (asia-east1): starting from n1-standard-4 - $0.253000
- Tokyo (asia-northeast1): starting from n1-standard-4 - $0.280600
- Seoul (asia-northeast3): starting from n1-standard-4 - $0.280600
- Belgium (europe-west1): starting from n1-standard-4 - $0.240580
- London (europe-west2): starting from n1-standard-4 - $0.281520
- Netherlands (europe-west4): starting from n1-standard-4 - $0.240580
Ray on Vertex AI
- Machine types - starting from $0.227999/hour
- Accelerators - starting from $0.864/hour
- Disks - starting from $0.048 per month, per GB
Vertex AI Workbench
- CPUs: starting from $0.045564 per vCPU/hour
- Memory: starting from $0.0061068 per GB hour
- Accelerators: starting from $0.492 per hour
- Disks: starting from $0.144 per GB per Month
Google Cloud Vertex AI Deals
Google Cloud Vertex AI Free Deals
Google Cloud Vertex AI Pricing Overview
Google Cloud Vertex AI offers a flexible and scalable pricing structure tailored to meet the diverse needs of businesses at different stages of their AI journey. Below is an overview of the pricing packages, free trial options, and what is included in each package.
Google Cloud Vertex AI Pricing Packages
Imagen Model for Image Generation
- Cost: $0.0001 per image
- Usage: Ideal for applications requiring image generation capabilities.
- Cloud: Charges are based on cloud usage.
Text, Chat, and Code Generation
- Cost: $0.0001 per 1,000 characters
- Usage: Suitable for generating text, chat responses, and code snippets.
- Cloud: Charges per character generated.
Text Data Upload, Training, Deployment, and Prediction
- Cost: $0.05 per hour
- Usage: Best for text data processing tasks, including training and deploying models and making predictions.
- Cloud: Hourly charge for cloud usage.
Video Data Training and Prediction
- Cost: $0.462 per node hour
- Usage: Designed for training and predicting with video data.
- Cloud: Charges based on node hours used.
Image Data Training, Deployment, and Prediction
- Cost: $1.375 per node hour
- Usage: For training models with image data, deploying them, and making predictions.
- Cloud: Node hour-based pricing.
Google Cloud Vertex AI Free Trial and Freemium Version
Google Cloud Vertex AI provides a generous free trial and freemium options:
- Free Trial: New customers receive $300 in free credits to explore and use Vertex AI services over 90 days.
- Freemium Version: Some basic features and services are available at no cost, enabling users to get started without an initial investment.
Included Features
Vertex AI offers a range of features across its pricing packages, making it versatile for various AI and ML needs:
- Data Management: Tools for uploading, storing, and managing large datasets.
- Model Training and Deployment: Comprehensive support for training and deploying machine learning models, including automated ML and custom model options.
- Prediction and Analysis: Capabilities for real-time and batch predictions, including robust analytical tools.
- Integration: Seamless integration with other Google Cloud services, ensuring a cohesive workflow across different cloud solutions.
- Support: Access to extensive documentation, community forums, and tutorials. Note that higher-tier support options, such as phone and live chat, may not be included in the free version but are available in paid packages.
Detailed Vertex AI Pricing for Specific Services
Vertex AI AutoML Models
- Training: $3.465 per node hour
- Training (Edge on-device model): $18.00 per node hour
- Deployment and Online Prediction: $1.375 – $2.002 per node hour
- Batch Prediction: $2.222 per node hour
Vertex AI Forecast and AutoML ARIMA+
- Prediction: $0.2 per 1K data points (0-1M points), $0.1 per 1K data points (1M-50M points), $0.02 per 1K data points (>50M points)
- Training: $21.25 per hour in all regions
Custom-Trained Models
Pricing varies based on machine type and configuration. For example, NVIDIA Tesla P100 costs $1.84 per hour, and NVIDIA Tesla T4 costs $0.4025 per hour.
Vertex Explainable AI
Charges are the same as prediction rates. However, explanations take longer, potentially increasing costs due to additional node usage.
Vector Search
- Index Serving: Varies by machine type and region, e.g., $0.094 per node hour for e2-standard-2 in us-central1.
- Index Building: $3.00 per GiB of data processed.
Vertex AI Feature Store
Online Operations: $0.1 per node hour for data ingestion, $0.38 per node hour for optimized online serving, $1.2 per node hour for Bigtable online serving.
Vertex AI Workbench
Instance Pricing: Varies by VM configuration and region. For instance, CPUs in Taiwan (asia-east1) cost $0.0303065 per vCPU/hour for E2 machine types.
Support Options
Vertex AI provides various support options depending on the package:
- Free Version: Includes access to email support, forums, knowledge bases, and video tutorials.
- Paid Version: Adds premium support options like phone and live chat support, ensuring prompt assistance for critical issues.
Additional Details
- Entry-level Setup Fee: There is an optional setup fee, depending on the specific requirements and services chosen.
- Customizable Pricing: Users can manage costs by allocating resources manually and using the pricing calculator to estimate expenses based on projected usage.
- Currency: Prices are listed in USD. If you pay in a currency other than USD, the prices in your currency on Cloud Platform SKUs apply.
- Billing Increments: Usage is billed in 30-second increments for training and prediction.
- Optimization Options: Includes optimized TensorFlow runtime, co-hosting models, and more ways to manage and reduce costs.
Google Cloud Vertex AI provides a comprehensive and flexible pricing structure, making it accessible for businesses of all sizes to leverage advanced AI capabilities. Whether you’re just starting with AI or scaling your operations, Vertex AI’s pricing plans are designed to provide value and support throughout your AI journey.