AI model selection interface showing different model options

Configuring AI Models

Nectir AI allows you to select different AI models for your Assistants, each with unique capabilities and performance characteristics.

Selecting an AI Model

  1. Navigate to your Assistant’s settings
  2. Click on Advanced in the settings sidebar
  3. Choose from the available models based on your Group needs:
    • GPT-4.1: OpenAI’s latest model with enhanced reasoning capabilities and improved instruction following. Excels in generating accurate, context-rich responses for healthcare, STEM, and social sciences.
    • Claude Sonnet 4: Anthropic’s advanced model with exceptional reasoning and instruction following. Shines in humanities, literature, and creative writing with its creativity and conversational flow.
    • Gemini 2.5 Pro: Google’s most advanced model with exceptional multimodal capabilities and reasoning. Excels in complex problem-solving, code generation, and handling diverse content types including images, audio, and video.
Which model should I choose?If you’re unsure which model to select:
  • GPT-4.1 is an excellent default choice for most educational use cases with strong overall performance
  • Claude Sonnet 4 excels in humanities, writing assistance, and generating more nuanced, detailed explanations
  • Gemini 2.5 Pro excels in multimodal content, complex reasoning, and technical subjects requiring code generation
  • For technical or STEM subjects, any model performs well

Comparing Model Capabilities

FeatureGPT-4.1Claude Sonnet 4Gemini 2.5 Pro
ReasoningStrong complex problem-solvingOutstanding analytical and critical thinkingExceptional complex reasoning and analysis
Response StyleDirect, concise, and preciseConversational, elaborative, and nuancedComprehensive, detailed, and analytical
Best ForHealthcare, STEM, Social SciencesHumanities, Literature, Creative WritingTechnical subjects, coding, multimodal content
StrengthsAccuracy, precision, problem-solvingCreativity, reasoning, ethical considerationsMultimodal capabilities, code generation, reasoning

Understanding AI Model Temperature

Temperature is a crucial setting that controls how creative or focused your AI Assistant’s responses are. Think of it as a dial that adjusts the balance between consistency and creativity in your Assistant’s answers.
AI model temperature scale showing the spectrum from focused to creative responses

How Temperature Works

Temperature controls how creative or focused your AI Assistant’s responses are. Rather than dealing with specific numbers, think of temperature as three distinct modes:
  • Predictable: Consistent, reliable responses ideal for factual learning
  • Balanced: Mix of consistency and creativity for general education
  • Creative: Varied, exploratory responses perfect for open-ended learning

Choosing the Right Temperature for Learning

Temperature Recommendations by Subject

Subject AreaRecommended ModeReasoning
MathematicsPredictableRequires precision and consistency in calculations
Sciences (STEM)Predictable to BalancedBalance of accuracy for facts with some explanatory flexibility
Medicine/HealthcarePredictableCritical information requires high accuracy
HistoryBalancedBenefits from varied perspectives while maintaining factual accuracy
LiteratureBalanced to CreativeEncourages diverse interpretations and creative analysis
PhilosophyBalanced to CreativeSupports exploration of different viewpoints and arguments
Creative WritingCreativeMaximizes creative potential and diverse approaches
BusinessBalancedBalances analytical precision with strategic creativity
Language LearningBalancedProvides varied examples while maintaining grammatical accuracy

Setting Temperature for Your Assistant

  1. Navigate to your Assistant’s Advanced settings
  2. Locate the Temperature setting
  3. Choose the mode that best fits your course requirements:
    • Predictable for subjects requiring accuracy and consistency
    • Balanced for general education with moderate variety
    • Creative for subjects encouraging exploration and creativity
  4. Test with sample questions to assess the response style
  5. Adjust as needed based on student feedback and learning outcomes
Start with Balanced mode if you’re unsure. This provides a good mix of consistency and creativity for most educational scenarios. You can always adjust based on how your students interact with the Assistant and the type of responses you want to encourage.

Model Literacy and Future-Proofing

Introducing different AI models helps build AI literacy skills. By engaging with large language models such as GPT-4.1, Claude Sonnet 4, and Gemini 2.5 Pro, students learn how diverse models yield different results, preparing them for an AI-driven workforce. You may want to encourage students to:
  • Compare model outputs for the same question
  • Analyze different response styles and approaches
  • Understand which model performs better for specific types of tasks

Tools

Configure which tools your Assistant can use to enhance its capabilities. Tools allow your Assistant to perform specialized actions beyond standard conversation, making it more powerful and adaptable to your needs.

Configuring Tools

  1. Navigate to your Assistant’s Advanced settings.
  2. Locate the Tools section.
  3. Turn on or off available tools based on your course or Group requirements.

Available Tools

Currently, Nectir offers the following tool:
  • Semantic Search: When enabled, Semantic Search helps your Assistant find the most relevant parts of your uploaded documents to answer each question. Without Semantic Search, your Assistant still uses your materials, but it relies on what it has learned in general, which can be less precise. With Semantic Search, answers are more likely to directly reference your actual documents, making responses clearer and more accurate for your users.

Next Steps


Video Tutorial

See Advanced Configuration in Action

Watch how to select AI models and configure temperature settings