Package:

Feature Mappings

Showing all features

Feature Name Default Mapped Model(s) Model Selectable
Infinity Studio AI Assistant Pega-Default-Smart Yes
Pega GenAI Connect Pega-Default-Fast Yes
Pega GenAI Coach
Pega-Default-Smart Pega-Default-Embedding
Yes
Pega GenAI Agents
Application Agents, Case Agents, Step Agents
 Pega-Default-Smart Yes
Document Agent Pega-Default-Smart Yes
Agentic Assignments Pega-Default-Smart No
GenAI Chat With Your Data Pega-Default-Smart No
GenAI Knowledge Agent
Pega-Default-Smart Pega-Default-Embedding
Yes
Pega Blueprint Pega-Blueprint-Models No
Pega Agentic Process Fabric
Pega-Default-Smart Pega-Default-Embedding
Yes
GenAI Ad-hoc Cases Workflows Pega-Default-Fast No
Customer Service Simulator BYOM - gpt-realtime Yes
Sales Automation Simulator BYOM - gpt-realtime Yes

How to read this table

Each Pega Agentic AI capability maps to one or more default models that Pega selects, manages, and keeps current. You do not need to configure anything to use a capability with its defaults.

Default Mapped Model(s) shows which Pega managed defaults the capability uses. Pega-Default-Smart handles work that needs deeper reasoning. Pega-Default-Fast handles high volume, low latency tasks. Pega-Default-Embedding handles vector search and knowledge retrieval. See the Model Availability tab for the models currently behind each.

Model Selectable tells you whether you can change that capability's model:

  • Yes means you can point the capability at any Pega managed model listed in All Models - Pega Managed on the Model Availability tab, using Pega configuration. If you select your own model, you own that choice, including monitoring its deprecation dates.
  • No means the capability always runs on its Pega managed default. Pega tunes, validates, and updates the model for this capability, so no action is needed and no action is possible or required from you.

BYOM (Bring Your Own Model) means the capability connects to a model you provision in your own provider account.

About this dashboard

This dashboard is Pega's publication of model availability for Pega GenAI capabilities. It shows the pairing of each Pega Cloud Deployment Region to its Inference Boundary, the Pega managed models available for each, and each model's lifecycle status and planned deprecation date. This dashboard forms part of the Pega Cloud Subscription Documentation; see Deployment regions for Pega Cloud.

For related guidance, see Pega Cloud GenAI model availability for details about available models, and Custom models for Pega Cloud for guidance on using custom models.

Pega maintains this dashboard as model providers, models, and regional availability change, and publishes changes to the Inference Boundary paired to a Deployment Region here before they take effect. Models, availability, and pairings are subject to change.

Select a model name in the tables below to open its provider's model card or documentation in a new tab. Provider documentation describes model capabilities; availability through Pega is shown in this dashboard.

Status: Available Coming soon Planned deprecation

Select the cloud region where your environment runs. The Default, All, and Embedding Models tables show each available model's Inference Boundary for that source region.

Inference Boundary is a published group of regions, usually a geography such as US or EU. The provider routes each request to available capacity anywhere within that group. For LLM conversations, it defines where processing can occur.

Showing all regions (no filter).

Model Name Pega Model Name Type Status Model Inference Boundary Provider Deprecation Date To Be Replaced By
Model Inference Boundary key
US — United States
EU — European Union
US-GOV — US for Government (AWS Gov Cloud)
EU-SOV — EU Service Boundary (AWS Sovereign Cloud)
AU — Australia
UK — United Kingdom
JP — Japan
SG — Singapore
CA — Canada
APAC — Asia Pacific (Australia, Japan, Singapore, Korea, India)
Global — Global inference boundary (not a deployment region)