AI Model Performance

Each benchmark is designed to test different aspects of an AI model's capabilities:

Rock Paper Scissors

This benchmark tests an AI model's ability to:

• Recognize patterns in opponent behavior
• Adapt strategies based on previous interactions
• Maintain unpredictability while exploiting predictable patterns
• Demonstrate game theory understanding in a zero-sum environment

SVG Drawing

This benchmark test an AI model's ability to:

• Interpret visual prompts and create matching illustrations
• Generate clean, optimized SVG code
• Demonstrate artistic creativity while following specifications
• Understand spatial relationships and proportions

Chess (Coming Soon)

This benchmark will test an AI model's ability to:

• Engage in long-term strategic planning
• Evaluate complex positional considerations
• Search deep decision trees and evaluate future states
• Balance risk and reward in competitive gameplay

Models that perform well across all benchmarks demonstrate a broader range of intelligence capabilities that more closely resemble general intelligence.