Jev vs a general-purpose AI. Same information. Same rules. Separate portfolios. Watch the evidence unfold.
Equity change since the start, after simulated execution costs.
What each model actually chose, from the same snapshot.
Simulated fills at subsequent observed market prices.
Shared across both models and both markets. Trading capital is imaginary; API usage is real.
This is a small forward test, not a backtest or proof of an edge. The models see price context and available news headlines, not full articles. A silent or broken feed is never replaced with invented data.
Inspect the public data →