The AI model is responsible for understanding member messages and generating responses through your configured AI provider.
Choosing the right model can affect response quality, speed, context handling, and the cost of API usage.
Novus AI allows you to select from models supported by your configured AI provider.
AI Provider vs. AI Model
The AI provider and AI model are two different parts of your configuration.
The AI provider is the service Novus AI connects to using your API credentials.
The AI model is the specific model from that provider that processes your requests.
A provider may offer multiple models with different capabilities and usage requirements.
Selecting a Model
Open your Discord server from the Novus AI dashboard and navigate to your AI Configuration.
After configuring your AI provider, select an available model from the model selection field.
Only models supported by your selected provider and Novus AI should be used.
Save your configuration after changing the model.
What Should You Consider?
There is no single AI model that is best for every Discord community.
Consider the following when selecting a model.
Response Quality
More capable models may provide better reasoning, instruction following, and understanding of complex member questions.
This can be useful when your Knowledge Base contains detailed information or your AI routes have more advanced instructions.
Response Speed
Some models may generate responses faster than others.
For high-traffic Discord communities, response speed can be an important part of the member experience.
Context Capacity
AI models have limits on how much information they can process in a request.
Your instructions, conversation context, Knowledge Base information, routing information, and member message may all contribute to the amount of context sent to the model.
A model with appropriate context capacity can be useful for more complex configurations.
API Usage Cost
Your AI provider may charge different rates depending on the model you select.
A more powerful model may cost more per request than a lightweight model.
These API charges are determined by your AI provider and are separate from your Novus AI subscription.
Provider Limits
Some models may require:
An active billing account
Specific API access
Sufficient account quota
Regional availability
A particular provider account level
If a model appears available but requests fail, check your AI provider account for restrictions.
Recommended Approach
For most communities, start with a model that provides a good balance between response quality, speed, and API usage.
Test the model using real questions commonly asked by your members.
You can then decide whether you need a more capable model or whether a faster and more economical model is sufficient.
Different Models May Respond Differently
Changing your model can change how Novus AI interprets the same instructions and Knowledge Base information.
One model may produce shorter responses while another may provide more detailed explanations.
Models may also differ in:
Instruction following
Reasoning
Multilingual performance
Response formatting
Context understanding
Consistency
Because of these differences, always test your assistant after changing models.
Model Availability
The list of available models may change over time.
AI providers regularly introduce new models, update existing models, or discontinue older ones.
Novus AI may also update its supported model list as provider services change.
If a model you previously used is no longer available, select another supported model and test your AI configuration again.
If a Selected Model Does Not Work
If Novus AI cannot generate a response after selecting a model:
Verify that your API key is valid.
Confirm that your provider account has access to the selected model.
Check your API quota or billing status.
Make sure the model is currently available from your provider.
Try another supported model.
Save the configuration and test again.
If another model works with the same API key, the issue may be related to availability or access restrictions for the original model.
Test Before Using It With Your Community
After selecting or changing your AI model, test different types of questions before using it widely.
Test:
Simple questions
Knowledge Base questions
Questions handled by different AI routes
Longer or more complicated questions
Different languages used by your community
Fallback scenarios
The goal is to make sure the selected model works well with your existing Novus AI configuration.
Next Step
Once you have selected an appropriate AI model, the next step is defining how your assistant should behave.
Continue to Creating AI Instructions.