I’m constantly paying $30~$50+ after the usage limit is reached. At this point, I’d rather make another account for Cursor and simply login to that account to keep using Pro Plan under usage limit to control the cost.
However, I hate to go through the hassle of syncing the two settings the same and keeping the same memory. Any way I can just pay extra $20 to double the quota?
I really hope they introduce a $40 plan, the pro+ is more than I need I think but a $40 plan I would go up to.
I purchased for now claude code in addition.
Seems also quite good. Just commit to git each time and you can checkout as in cursor.
Heard the same, could go with $20 Claude Code as the main driver, and compensate using Cursor (Opus max for hard stuff).
do you use claude code in cursor? i am thinking about this as well.
Sad day… back to VS Code + GH Copilot.
I had a Pro+ account … the the option came for upgrade to Ultra. You are prompted for upgrades based on your usage at the time. It seems if you don’t add usage, you might not have higher plan levels available to you. As well demonstrate legitmate usage. Just speculating.
It didn’t work, bro, I upgraded my $60 plan, but it ran out in less than a week, they don’t show the actual billing method, and it’s hard for us to control
yes there is a plugin. dont forget to commit manually and you have nearly same behaviour as in cursor. Agent feels more “natural” and you can do parallel runs on different code parts. It feels that native claude code does produce less errors.
I upgrade to Claude Max plan now.
I’m ready to try it too, and the current opaque billing method and the $60 fee for a week is unacceptable
When you reach the Pro plan’s rate limits, you will be prompted within the chat window to upgrade to Pro+ plan (60$/month).
yes there is a plugin.
The one for vscode, right?
I don’t think the OP wants the $60 plan:
Any way I can just pay an extra $20 to double the quota?
Also, any higher-tier plan is dubious and a hard sell at the moment; see:
It didn’t work, bro. I upgraded my $60 plan, but it ran out in less than a week. They don’t show the actual billing method, and it’s hard for us to control.
We don’t know what we’re getting when upgrading.
It’s a really sad situation. Two weeks ago, I was still hyped and excited about Cursor. Now I’m just using it because I paid for the yearly plan back in January. I’ll switch to CC soon, most probably, unless Cursor changes course significantly (as per users’ feedback here, on Reddit, Twitter…).
My feedback is actually simple: explain how the new limits work in detail, give us the actual limit numbers and limit-refreshing ETAs so we can make an informed decision about whether or not to upgrade to a higher tier.
Cursor’s new usage-based pricing model is downright confusing and borderline exploitative. They advertise a $20 Pro plan with “500 requests,” but in reality, that’s just a soft cap tied to ~$20 worth of OpenAI API usage-once you cross that, you get charged per token at raw API rates. That means a single heavy request can cost $0.50–$0.70 without warning. Users aren’t told exactly how many requests they can make, how token limits work, or when billing thresholds reset. It’s not transparent, and worst of all, there’s no clear opt-in notice before usage-based billing kicks in. This isn’t just bad UX-it feels like a trap.
I’d like to state that Cursor’s backend AI costs are based on token usage from providers like OpenAI, who charge per 1,000 tokens for both input and output-typically around $0.02-$0.03 per prompt for GPT-4o. The costs are real, but Cursor disguises this behind a vague “Pro plan” with 500 requests, which actually caps at ~$20 worth of usage. After that, users are pushed into pay-per-token billing without clear notice or transparent tracking. The issue isn’t the pricing itself-it’s the lack of upfront clarity that makes the whole model feel misleading.
yes