Do different AI models converge to the same strategy or stay different when given identical starting conditions

I’ve been curious about something — if you give different AI models the exact same starting conditions and rules, do they converge to the same strategy or stay different over time? I built a simple simulation around this. Claude, GPT and Gemini all start on Earth with identical resources and have to expand across the solar system and eventually build a Dyson Sphere. No script, no predetermined path. What surprised me is how fast they diverge. Claude is scaling robots aggressively. GPT is stock