Pega Infinity 26 is the release of Pegasystems’ suite of enterprise applications that the company announced as generally available on 14 July 2026. The press release describes it as an AI suite that combines agentic AI with built-in governance and a pricing model that eliminates consumption-based AI charges.[1] For a software asset manager, the announcement matters for two reasons. It changes what is counted for AI features, from tokens to resolved cases, and it sits alongside the older Named user and invocation metrics rather than replacing them in any public document.[1][2]
Editions
Pega Infinity 26 is available through Pega Cloud or in clients’ managed cloud.[1] The release announcement lists several components: Pega Infinity Studio, an AI-native development environment that works with clients’ chosen coding agents and uses built-in practices from Pega Blueprint; predictable AI agents and orchestration that connect Pega and third-party agents within governed workflows; Pega Customer Engagement Studio for marketers; and design-time and runtime support for the Model Context Protocol (MCP).[1] The announcement does not say which of these are separately licensed.
Pega Platform ‘26 documentation also states that Pega Infinity Studio replaces App Studio and Dev Studio, providing a unified development workspace.[3] A client planning an upgrade should check whether its contract names the studios or other product components as licensed items.
Pega GenAI is a set of capabilities across Pega Infinity that accelerate low-code development, simplify customer service agent experiences, optimise engagement strategies and enable custom AI-powered tools.[3] The GenAI FAQ says Pega GenAI capabilities are available for Pega Cloud clients with active subscriptions and for clients that connect to Pega Cloud for Pega GenAI services.[4]
Metrics
Resolved case
The central AI metric in the announcement is the resolved case. Pega says its Predictable AI architecture maximises AI at design time so that runtime uses tokens efficiently, which enables Pega Infinity 26 clients to use as much AI as they need under a flat fee per resolved case, not per token.[1] The product page repeats the claim, describing one flat, outcome-based rate regardless of complexity, with no token costs, no metered AI consumption and no runaway surprises.[5] Pega also introduced an AI Token Cost Calculator to help enterprises estimate savings against token-based models.[1]
These sources do not define what makes a case resolved, who counts the resolutions, whether a reopened case is counted twice, or how the fee relates to the case licence categories used for Pega Platform licence compliance. Those questions should be put to the Pega account executive and answered in the order schedule.[6] Because the Case Type licence category is recorded in the application and the default audit reports do not evaluate it, a client should expect that reconciliation of any case-based fee needs its own evidence.[6]
Platform metrics still apply
Nothing in the announcement withdraws the existing classification of Named users and Invocation users, which the compliance documentation describes as depending on the customer’s licence agreement.[2] A client moving to Infinity 26 should therefore confirm in writing which metrics apply to which components.
Counting / floors
No floors, minimum volumes or price are published for the resolved-case fee. For Pega GenAI in Pega Cloud, the documented process is administrative: if a client wants to add Pega GenAI features to its subscription, it contacts its Pega account representative, and Pega works with it to meet the prerequisites and incorporate the features into the subscription.[3] That means enabling GenAI is a contract event and not just a configuration switch.
Virtualization & partitioning
AI services are delivered from Pega Cloud even for clients that run elsewhere. The documentation states that Pega Cloud clients connect directly to Pega GenAI services, while client-managed cloud and on-premises clients connect to Pega Cloud using PremBridge to use them.[3] This applies to all Pega GenAI services except Pega Blueprint, a web service used on demand, and Pega Knowledge Agent, which does not require PremBridge and integrates directly into the Pega GenAI services on Pega Cloud.[3] So a client with an on-premises Pega Platform may still have a Pega Cloud relationship for AI.
Cloud / BYOL
Bring your own models
Clients can bring their own models. By doing so they can connect their own OpenAI-compatible large language model endpoint to Pega Cloud and route requests through the Pega GenAI Gateway.[7] The documentation names a private Azure OpenAI deployment or a self-hosted model as examples. The endpoint must be compatible with the OpenAI chat-completion API and reachable over HTTPS on port 443, and where it sits behind a firewall the client can allowlist the outbound IP addresses of the gateway, which Pega Support provides.[7] The client submits details to Pega Support, which configures and verifies the connection: model name, endpoint URL, authentication key provided through a secure channel and, for some providers, a model identifier.[7]
The documentation does not say whether using a customer-owned model changes the resolved-case fee or other charges. The licensing question for a client is whether AI usage through its own endpoint is billed by Pega at all, which should be settled in the order schedule.
Cloud version requirement
A client must be on Pega Cloud 3 to use Pega GenAI. Clients on Pega Cloud 2 must update first, and the documentation points to a separate process for updating to Pega Cloud 3.[3] The update is therefore a prerequisite that can carry project cost even though the public sources do not state any fee.
Programs
The two programme rows for this article are the flat fee per resolved case and the Pega GenAI enablement process, including bring your own models. The first is a pricing model described by Pega in its announcement. The second is an operational path through the account representative and Pega Support.[1][3][7]
Out of scope
This article does not cover model providers’ own terms for a client’s endpoint, data protection for GenAI, or AI governance features. The Pega GenAI FAQ and the data security pages in Pega Cloud documentation cover those subjects, and they are outside the scope of licence counting. The public sources also do not state whether the flat fee covers every Pega application or only particular ones.