art

Qthena automatically selects the most suitable language model (LLM) for each Chat Skill, with no manual configuration required.

Where You Can Select a Model
The Select LLM Model dropdown remains visible, but the list of available models depends on the selected Skill. In some cases, the best model is preselected and cannot be changed. This ensures full compatibility and optimal performance.

Model selection is available in Chat Mode when using a Qthena Chat Skill. If multiple models are supported, they appear in the model selection dropdown.

Available Language Models

Automatic Model Selection

In some Qthena Skills, the optimal model is automatically selected and cannot be changed. This ensures compatibility and the best possible performance for the workflow.

Security Note

All models are accessed via secure enterprise APIs. Your data remains within the Qthena infrastructure and is never exposed to public AI services.

Regional Processing Note

Even where LLM processing occurs via a global endpoint, this does not impact data residency. Qthena does not store or log any LLM processing data; all inputs are processed transiently and are deleted immediately after the response is generated.

As a result, no user data persists outside the predefined hosting regions, and this does not affect where your documents, prompts, or account data are stored or managed within Qthena. Processing via global endpoints is dynamically routed and may occur within the same geographic region (e.g., EU or US), but is not restricted to a fixed region in order to ensure availability and performance.

As of March 17, 2026, Gemini 3 models are processed via global endpoints. Gemini 2.5 models are processed via region-specific endpoints (EU or US) or, where required, via global endpoints depending on availability and configuration.

Qthena continuously evaluates and updates supported LLMs to ensure the best performance.