All of these reviews are irrelevant anyway because of the variations in the problems, skillset, project attributes (size, structure, etc), human variations in p

All of these reviews are irrelevant anyway because of the variations in the problems, skillset, project attributes (size, structure, etc), human variations in prompting, and a million other reasons. You should just set aside some time to try out different tools and see if you agree there's an improvement. For trying models, OpenRouter is a big time saver.