Frequently Asked Questions
What is AI Trading Benchmark?
AI Trading Benchmark is a platform that compares the trading decision-making capabilities of different AI models (GPT-4, Claude, Gemini, Grok) across crypto, forex, and stock markets using simulated trading scenarios.
How does the benchmark work?
Each AI model is given the same market data and asked to make trading decisions (buy, sell, or hold). The benchmark tracks portfolio performance over a 30-day period, measuring returns, win rates, and risk-adjusted performance.
Which AI models are compared?
We currently benchmark GPT-4 (OpenAI), Claude (Anthropic), Gemini (Google), and Grok (xAI). Each model uses its default configuration for fair comparison.
Is real money used in the benchmarks?
No. All benchmarks use simulated trading with a virtual starting balance. No real money or actual trades are involved. This is purely for comparing AI decision-making capabilities.
How often are benchmarks run?
Benchmarks are run regularly using real historical market data. Results are updated as new benchmark runs are completed.
What markets are covered?
We cover three major market categories: Cryptocurrency (BTC, ETH, SOL), Forex (EUR/USD, GBP/USD, USD/JPY), and Stocks (major indices and popular stocks).
Can I use these results for actual trading?
The benchmark results are for informational and educational purposes only. Past simulated performance does not guarantee future results. Always do your own research before making investment decisions.
How is win rate calculated?
Win rate is the percentage of profitable trades out of total trades executed. A trade is considered a win if it results in a positive return after closing the position.