- Getting started
- Data security and compliance
- Organizations
- Authentication and security
- Licensing
- About licensing
- Unified Pricing: Licensing plan framework
- Activating your Enterprise license
- Migrate from Test Suite to Test Cloud
- License migration
- Assigning licenses to tenants
- Assigning user licenses
- Deallocating user licenses
- Monitoring license allocation
- License overallocation
- Licensing notifications
- User license management
- Tenants and services
- Accounts and roles
- AI Trust Layer
- About AI Trust Layer
- Managing models in Model hub
- External applications
- Notifications
- Logging
- Data Export
- Testing in your organization
- Troubleshooting
- Migrating to Test Cloud
Configure model routing, replace models including Bring Your Own Model, and review configuration history for a product and feature in Model hub.
From the Model hub tenant grid, you can change how a product and feature routes LLM traffic, replace the model that backs it, and review its configuration history. All actions apply to the current tenant.
Prerequisites
- Access to Model hub at the tenant level.
- The product and feature are enabled for the tenant.
- For a custom LLM configuration — Bring Your Own Subscription (BYOS) or Bring Your Own Model (BYOM) — a configured endpoint or subscription. For setup, see Configuring LLMs.
Each procedure starts from the tenant grid. To open it, go to Test Cloud Admin > AI Trust Layer > Model hub, then select a tenant from the Tenants panel, or select View details on a tenant card. In the grid, locate the product and feature row to work with, and open its options menu to access the available actions.
Selectable and non-selectable products
Not all products support direct model selection, which affects how you configure routing and replace models:
- Selectable products, such as Agents and GenAI Activities, let you select the model directly.
- Non-selectable products, such as Test Manager, do not expose direct model selection. For these, the grid supports viewing and policy-based outcomes, plus any configuration options UiPath supports for that product.
Configure model routing
Configure changes where a product and feature routes LLM traffic and, where the product supports it, which model backs the mapping. To use your own subscription instead, see Replace a model.
- Open the row's options menu and select Configure model.
- Under Select model routing, select an available region. The region in use is labeled Current.
- Under Select model, select one of the models available in that region, if the product supports model selection.
- Select Save configuration.
The mapping's routing region, and its model where you changed it, reflect your selection.
Grayed-out regions have no compatible routing option for this product and feature, so no eligible model exists there. A disabled region does not mean UiPath does not support that region.
Replace a model
How you replace the model backing a mapping depends on whether the replacement is UiPath-managed or a custom LLM configuration.
Switch to a different UiPath-managed model
- Open the row's options menu and select Configure model.
- Under Select model, select the replacement model. When a model is deprecating, the dialog shows its deprecation date and the designated replacement model.
- Select Save configuration.
Switch to a custom configuration
To use a custom LLM configuration (BYOS or BYOM):
- Open the row's options menu and select Add custom model.
- Set up the custom LLM configuration. For details, see Configuring LLMs.
After you apply a custom replacement, your model appears in the Model hub list with Managed by set to Custom, and can be reused as a replacement option elsewhere.
You are responsible for the lifecycle of any custom LLM configuration, including migrating to a replacement before a model is deprecated. See Custom LLM configurations. For upcoming deprecation dates, see the LLM model deprecation timeline.
Upgrade a deprecating custom model
When a model you use through a custom configuration is deprecating, for example GPT-4, migrate before the deprecation date:
- Provision a replacement model, for example GPT-5, on your subscription.
- Update the custom LLM configuration to use the replacement.
- Validate and save the configuration. For details, see Configuring LLMs.
Revert to the default configuration
Model hub distinguishes between two configuration states:
- Initial configuration, labeled Default, is the state before any custom configuration is applied.
- Current configuration, labeled Edited, is the active configuration, based on the most recent change.
How you revert depends on the current configuration:
- If the mapping uses a UiPath-managed model, use Configure model to set the routing model and region back to the default values.
- If the mapping uses a custom LLM configuration, remove the configuration to return to the UiPath-managed default.
To compare the current values against the defaults, use the Columns control to add the optional default columns, such as Default model, to the grid.
View configuration history
View History shows the audit-style history for a mapping, so you can see how it changed over time and whether a user or the system made each change.
- Open the row's options menu and select View History.
For the selected mapping, the history shows how many times the model changed and, for each change, the following fields:
| Field | Example |
|---|---|
| Timestamp | 2026-06-05 13:42 UTC |
| Actor | System, or a user identity |
| Change type | Region changed / Model changed / BYOM assigned / BYOM removed |
| Before and after | GPT-4 (US) > GPT-5 (EU) |
| Reason (optional) | Deprecation replacement / Manual update |
A system action reflects an automated update, such as a UiPath-managed default change, while a user action reflects a manual configuration by an administrator.