Claude never had competition
And finding that out costed a lot of money until I just gave up.
Determining what is the best model for certain type of development can take up from minimal a week to a month.
Something that people did not check, EVERY TIME is that Claude actually does code maintenance. All models in exception of Claude did do that very, very bad actually.
For example, take the latest hype, kimi or whatnot and create something with it. Let claude check it afterwards. Yeah, it is that bad.
I also do not like that the most mainstream model is alledgedly and based on my own experience that i rely on of course, is the best. But don't worry, there is a new compatitor and yes, i cancelled Claude for this month and do not regret it at all so far. I am already two months using Meta Muse and I love it while it is probably not better.