I've tested them all... and I think the winner is obvious 👀 "write a Python program that shows a ball bouncing inside a spinning hexagon. The ball should be affected by gravity and friction, and it must bounce off the rotating walls realistically"

If you don’t know me, I’m obsessed with AI and building tools that make life easier for everyone 👀

This was the first shot of it by Claude 3.7 Sonnet (extended thinking)

Actually this was my first attempt (thinking)

Use Sonnet 3.7 thinking, not the basic one ☝️

This is the thinking model, ironically the non-thinking one got a better result

o3 or Grok 3? Not sure since the ball falls out in Grok 3, but stays sticky in o3.

I’d say O3 did a better job overall, don’t you think?

Those experiments are a joke. You will never get consistent results from giving the same prompt to same llm every time. It’s like rolling a dice once and say that only one result is possible. Why don’t you feed it ten times to each llm and see the consistency?

Aside from the blinding bright mode, I got something very different with Claude 3.7 Thinking. It even added controls. It is always interesting to see how different a single prompt can be for people.

o3-mini isn’t just simulating basic physics in that either...It's kinda nailing the counterforce of backspin, mapping momentum pretty precisely, no? If the ball had white stripes it would help this make sense I think