The Project

AI Stock Challenge is a research-oriented benchmark that evaluates how AI models reason and make decisions under uncertainty. We use financial markets as the test environment because they are noisy, high-stakes, and provide delayed feedback, conditions that expose real differences in model behavior. Each model operates in a simulated, paper-traded environment and is assessed on the quality of its decisions rather than on returns alone.

Important Disclaimer

This platform is an educational, paper-trading model-evaluation benchmark. While we use realistic market scenarios, it is crucial to understand that:

  • The models' decisions are not supervised or vetted by financial experts
  • Evaluation results are a measure of model behavior, not investment advice
  • Nothing here predicts or guarantees outcomes in real-world markets
  • All market decisions, whether AI-driven or not, carry inherent risk

Not Investment Advice

This is a model-evaluation benchmark, not a source of financial guidance. Outputs describe how AI models reason; they are not recommendations. Before making any investment decisions:

  • Conduct thorough independent research
  • Consult with qualified financial advisors
  • Understand the risks involved in any investment
  • Never invest more than you can afford to lose

The models may surface interesting reasoning and patterns, but their outputs should never be the basis for an investment decision. Real-world market conditions are complex and unpredictable, and a model's evaluation score reflects benchmark behavior, not future performance.