Choose Model
Overview#
Choose Model lets you select the large language model powering your agent. Pick from GPT, Gemini, or other supported options depending on the response quality, latency, and cost you require.
When to Use It#
- Creating a new agent and deciding which model should handle conversations
- Updating an existing agent to take advantage of a new model release
- Aligning the model choice with organization-wide compliance or pricing requirements
Key Benefits#
- Performance control — select models optimized for speed, cost, or advanced reasoning
- Flexibility — switch models without rebuilding the entire agent
- Transparency — understand key traits like context length and pricing before committing
UI/UX Specification#

Component Reference#
Component | Type | Description |
---|---|---|
Model dropdown | Select menu | Lists available models with descriptions |
Description panel | Text area | Displays details such as recommended use cases |
Save button | Primary button | Applies the selected model to the agent |
Cancel button | Secondary button | Closes the modal or reverts if you exit without saving |
How to Use#
- Open the agent configuration form (
Create Agent
orEdit Agent
). - Locate the
Model
dropdown within the General or Settings section. - Expand the list and review the available models. Pick the one that fits your needs.
- Save the agent settings to apply the change. The new model becomes active immediately after saving.
Notes & Tips#
- Consider token limits and cost implications when selecting models for high-volume chatbots.
- Some models may be restricted by plan or require additional approval—check the tooltip or plan documentation.
- After switching models, test the agent's responses to confirm the tone and accuracy still align with your brand.
FAQ#
Can I test a model before committing?#
Use the agent preview or staging environment to validate responses. If the model underperforms, switch back at any time.
Does changing models retrain my data?#
No. Your knowledge base stays intact. Model changes affect how the agent interprets prompts, not the underlying data.
Are all models available in every region?#
Availability can vary. The dropdown only lists models enabled for your organization and region.
Last updated on