Have you ever wondered the difference in many of the commonly used AI tools out there? Many people just use ChatGPT because it’s the most popular out of the bunch, but there are reasons to switch up your AI usage in some cases. In this article, you will see some comparisons, with AI generated text and picking it all apart, so why not test something it’s most commonly used for?

CODING

Now, coding is a big part of AI after its introduction into the field because it makes troubleshooting much easier and just saves time in the starting stages of making a website, game, or government database. (There is evidence that AI generated code was used in the making of some vital government systems, take that as you will.) But let’s test out a simple prompt, but unique enough to not just be copy and paste from training data. Let’s make a game!
Prompt: Write me a top down shooter with walls, enemies, and a procedurally generated map. Make it in Python or Pygame

ChatGPT 04-mini (Reasoning Mode): To be honest, it’s not a bad attempt, but this being ChatGPT’s smartest model, I’d kinda expect a little more. There’s no dying to the enemies and the enemies move around randomly, but there is collision with the walls and it is procedurally generated, so I guess?

Claude 3.7 Sonnet: Ok, this blew me away, this is the base model with barely ANY reasoning, zero forethought whatsoever and it makes a game with so many more features than I expected. The map is procedural, there are HEALTH BARS, intelligent enemies, moving camera, this is amazing. Coding? Blows me away.

Gemini 2.5-Pro-Exp: This is a little less impressive than Claude, but much better than ChatGPT by far. It has some features from Claude but lacks being as dynamic. Still, I guess it did fit the parameters.

Let me know what you think of the results! Personally I think Claude won by far and ChatGPT was on the bottom. It was super disappointing to not see more of a fight, but what can you do?

Stay tuned for more AI comparisons!