This site is currently a counterproductive trash that will simply waste your time on using it. There is some promise in LLM comparison but the limits of this make it utterly useless, even as a litmus test, since you can't check anything practical with it.

It has a very restrictive one-time prompt that stops working pretty fast sending you to buy their subscription and there is even no way to ask the models to fix or continue printing the unfinished code since the input field gets locked right after the inital prompt.