Skip to main content

Model Catalog

The Model Catalog serves as a centralized repository for discovering, evaluating, and comparing all available Large Language Models (LLMs) on the CAIP platform. It enables users to quickly identify models that meet their business and technical requirements by providing detailed information about capabilities, pricing, supported modalities, deployment regions, and lifecycle status.

The Model Catalog is designed to help users:

  • Browse available foundation models and agents.
  • Compare models across providers.
  • Understand model capabilities and limitations.
  • Review pricing and context window sizes.
  • Evaluate supported modalities and features.
  • Track model lifecycle and retirement dates.
  • Select the most appropriate model for specific use cases.

Model Catalog Model Catalog Overview

Model Comparison Model Comparison View


Search and Filtering

Users can narrow down available models using the following filters:

FilterDescription
SearchSearch models by name or provider
ProviderFilter models by vendor (Azure, OpenAI, Anthropic, etc.)
RegionView models available in specific deployment regions
ModesFilter by supported model interaction types
Expand AllExpands all model cards to display detailed capabilities

Model Information

FieldDescriptionExample
Model ProviderIndicates the cloud or model providerAzure
Model NameThe unique model identifierGPT-4o
Capability IndicatorsIcons shown beside the model name represent supported capabilities
Interaction ModeIndicates supported interaction patternsCHAT
Region AvailabilityShows where the model is hosted and availableEU
Retirement DateDisplays the planned retirement date of the model2026-10-01

The Retirement Date provides visibility into the model lifecycle and helps customers plan migrations to successor models before the model becomes unavailable.


Retirement Management

Importance of Retirement Dates

Retirement dates are highlighted to ensure users have sufficient visibility into upcoming model deprecations. When a model reaches its retirement date:

  • New deployments may no longer be permitted.
  • Existing integrations may stop receiving support.
  • Security updates and enhancements may cease.
  • Users should migrate workloads to recommended successor models.

Best Practice

Users are encouraged to:

  • Regularly review retirement dates in the Model Catalog.
  • Plan migrations at least several months before retirement.
  • Test replacement models in lower environments.
  • Update dependent agents, applications, and workflows before model decommissioning.

The Retirement field serves as an early warning mechanism, enabling proactive lifecycle management and uninterrupted service continuity.


Permission Levels

See the centralized Permissions Matrix for owner/member Create, Read, Update, and Delete permissions across all portal features.


Learn More

For the full list of available models and usage details, see the Model Catalogue documentation.