In the gym comparison you see them answer the same prompt side by side. Here is why they answer differently: who makes them, how they are built, where they are strong and where they fall.
The AIs, one by one
Claude (Anthropic, United States): Writes long without losing the thread, and says when it does not know
ChatGPT (OpenAI, United States): The best known, and the one that sounds certain even when it is wrong
Gemini (Google, United States): The one that lives inside Google, with the roomiest memory
Qwen (Alibaba Cloud, served by Groq): The open one, the one you can take home
Mistral (Mistral AI, France): The European one: fast, light, and with the data staying here
Grok (xAI, United States): Wired into X, with the loosest tongue of them all
Perplexity (Perplexity AI, United States): Answers with the sources at the bottom, so you can check it