Documentation

🤖Model Selection Guidance

Use Consensus to get informed recommendations for model selection and profile optimization

← Back to Documentation

Model Selection Guidance

Leverage collaborative consensus to make informed decisions about which AI models to use for specific tasks and how to optimize your custom profiles.

How to Get Model Recommendations

Use the Consensus feature to ask questions about model selection. Multiple AI models will discuss and provide validated recommendations.

Example Questions

  • "Which models are best for Python code generation?"
  • "What's the most cost-effective model for API documentation?"
  • "Compare GPT-4 and Claude Sonnet for code review"
  • "Which models have the largest context windows?"

Selection Criteria

Consider these factors when selecting models for your custom consensus profiles:

Performance Factors

  • Task-specific capabilities (coding, reasoning, writing)
  • Context window size
  • Response quality and accuracy
  • Specialized domain knowledge

Practical Considerations

  • Cost per request
  • Response latency
  • Provider reliability
  • Model availability

Building Balanced Profiles

Create effective consensus profiles by combining models with complementary strengths.

Diversity Strategy

Select models from different providers to reduce correlated errors and bias. For example: OpenAI + Anthropic + Google + Meta.

Cost Balance

Mix premium and budget models based on role importance. Use stronger models for the curator, cost-effective models for initial discussion.

Task Optimization

Match model capabilities to task requirements. Use coding specialists for development tasks, reasoning models for architecture decisions.

Pre-Built Templates

Start with professionally optimized templates that balance performance, cost, and quality for common use cases.

Speed-Optimized

Fast models for rapid prototyping and iterative development

Quality-Focused

Premium models for critical decisions and production code

Cost-Efficient

Budget-conscious selections that maintain quality

Specialized Tasks

Task-specific configurations for security, ML, debugging

Getting Started

Use Consensus to ask about model selection, then create custom profiles in Settings.

Download Hive Consensus