The $20/month 500 “requests” lasted me about 15 hours of coding. I’m trying to have a swag at how much an additional 15 hours of coding would cost me if I buy into MAX.
I know all the caveats… just wondered if someone had already done this and what your experience might have been.
I love cursor and I think it’s better than windsurf, but I don’t love the pricing. It seems like whoever is coming up with the names for the new openAI models is also coming up with the pricing for cursor. I’m sure it makes sense to someone with a lot more brain power than me, but for me and the rest of us normies we need something that doesn’t require a math degree to understand. 
If I were you (and I’ve been you) I would just switch to trying out some other vibe coding option until your credits refresh, otherwise be prepared for some interesting bills! I burned through about 400 credits last night just working on my personal website.
Don’t know if this helps, but I’ve been going hard on usage and in particular with this new max mode. I just had to increase my budget again. It does start to add up.
I do feel like it has helped me make progress with my project, which is pretty big. But sometimes it feels like the models are just smart enough to do a little but too clumsy to do it right - and I end up wasting more tokens to fix annoying typescript issues or catch things that are not what I specifically asked for.
Also, I just tried using normal gemini-2.5-pro-preview-05-06 for a task, it went ahead and wasted paid usage and then just dipped. My suspicion is that it just did not have enough context to finish and quietly quit on me.
I tried again with the same model but on MAX mode - it was able to finish and give me the result I was expecting.
