AI extensions
This sections lists extensions to the LLM integration in ConSol CM.
Support for reasoning models (#669143)
ConSol CM now supports AI models with reasoning capabilities, such as GPT 5.5, Claude Opus 4.7, or Gemini 3.1. The system automatically handles the necessary token configuration for these models to ensure stable API communication.
Two new settings have been added to the LLM connections page of the Web Admin Suite to configure reasoning capabilities:
- Reasoning enabled: Set to true when using a model with reasoning capabilities. The default value is false.
- Reasoning effort: Specifies the computational effort for requests. Higher effort levels consume more tokens. Available values (e.g., low, medium, high, xhigh) depend on your model provider. Refer to your provider's documentation for supported options.
The Reasoning effort setting is only displayed when Reasoning enabled is set to true. The reasoning options remain hidden for Ollama provider types, as reasoning features are not supported in this integration.
Reasoning models typically require more tokens than standard models. If you experience request failures related to token limits, please increase the Token limit setting in the LLM connection configuration.
Support for Anthropic models (#669239)
It is now possible to directly use Anthropic models in ConSol CM. The option Anthropic has been added to the Provider type selector on LLM connections page of the Web Admin Suite for this purpose.