Vertex AI quotas and limits

Google Cloud uses quotas to help ensure fairness and reduce spikes in resource use and availability. A quota restricts how much of a Google Cloud resource your Google Cloud project can use. Quotas apply to a range of resource types, including hardware, software, and network components. For example, quotas can restrict the number of API calls to a service, the number of load balancers used concurrently by your project, or the number of projects that you can create. Quotas protect the community of Google Cloud users by preventing the overloading of services. Quotas also help you to manage your own Google Cloud resources.

The Cloud Quotas system does the following:

• Monitors your consumption of Google Cloud products and services
• Restricts your consumption of those resources
• Provides a way to request changes to the quota value

In most cases, when you attempt to consume more of a resource than its quota allows, the system blocks access to the resource, and the task that you're trying to perform fails.

Quotas generally apply at the Google Cloud project level. Your use of a resource in one project doesn't affect your available quota in another project. Within a Google Cloud project, quotas are shared across all applications and IP addresses.

There are also limits on Vertex AI resources. These limits are unrelated to the quota system. Limits can't be changed.

Request quotas

The following quotas apply to Vertex AI requests for a given project and supported region. For example, in a single project, you can have up to 30,000 online prediction requests per minute in one region and another 30,000 online prediction requests per minute in another supported region.

¹Resource management requests include any request that isn't a job, an LRO, an online prediction request, a Vertex AI Vizier request, an ML metadata request, a Vertex AI TensorBoard Timeseries Insights API read request, a Vertex AI Feature Store request, a Vertex AI Feature Store streaming request, or a Vector Search request.

² This quota applies for public endpoints only. Private endpoints have unlimited requests per minute.

Jobs or LROs include the following requests:

• Create or delete a dataset.
• Import or export data to or from a dataset.
• Create an endpoint.
• Create or delete a custom job.
• Create or delete a data-labeling job.
• Create or delete a hyper-parameter tuning job.
• Create or delete a batch-prediction job.
• Create or delete a model.
• Upload, delete, or export a model.
• Create or delete a notebook-runtime template.
• Assign, delete, start, or upgrade a notebook runtime.
• Create, delete, or update a model monitor.
• Create or delete a model monitoring job.

For quota information for Generative AI models, see Generative AI on Vertex AI quotas and limits.

AutoML model quotas

The following quotas apply to each data type and objective for a given project and region. For example, in a particular project and region, you can deploy 10 AutoML image classification models and 10 AutoML image object detection models for a total of 20 deployed models.

AutoML model limits

The following limits apply to each data type for a given project and region. For example, in a particular project and region, you can include a maximum of 1,000,000 images in a batch request input.

Custom-trained model quotas

The following quotas apply to Vertex AI custom-trained models for a given project and region.

Training