I have a business subscription which is worth $40/month and I’m really questioning if it’s worth. Even for the $20/month functionality, It’d wouldn’t make sense.
All of the recent updates being pushed (mainly the agent functionality) are remotely not there in terms of reasonable quality. Functionality is crappy and slow.
I wonder if these are being tested enough before releasing. Or rather they are just trying to move too fast and let things break.
What version of cursor would make it worth $200/mo to you?
That’s the question you really should be asking
They could have priced it $200/mo. 
More over it’s really not about the version. Because of a lot of recent issues and continuous updates being pushed. I’m questioning the price to value ratio.
As for the version I’m on 0.46.9 which is the latest and still have issues with the agent not responding correctly. Responses are totally misformatted. 
See?
People would pay the money necessary to make this a real company what is this 20 employees ■■■■■■■■
This is my question as well. The GUI gets more bugs every update, the model is switching from good external models to internal dumb model, the price is increasing.
My year subscription is ending and I think I’ll abandon Cursor.
I’m not willing to have a token based bill, because 90% of the answer I get are wrong, and I’m not willing to pay for wrong answer.
Not just almost ok answer, but often agents break the code, made me waste hours looping into messy code.
Really, why should I pay for broken code, for mistakes, wrong answers?
Agents NEVER follows rules, when tries to explain what it did, it is barely understandable.
When 90% of the code produced WILL be good, I can stand paying for 10% of mistakes, but right now it is the vice versa. Now is 90$ per mistakes and 10$ for ok code.
I don’t think it’s worth it.
Totally agree
Spending 75% of my time fixing errors caused by cursor