This tier list of AI models is probably more accurate than most official benchmarks. Because it is compiled not based on the results of test tasks, but on the basis of real use for money. What do you agree with and what do you disagree with? 😰