Not sure where to find that, but I did notice on Auto for model it always applies usage toward Grok/Composer. However in Grok Dashboard usage it is saying it used Claude Opus, this may be part of why usage is climbing so fast?
I wanted to share my experience with the new plans, as feedback from someone who uses Cursor for coding every day.
I like the idea of having Grok, Grok Bot, and Cursor in one subscription, and I understand that a single usage limit for everything is part of that vision.
On Cursor Pro ($20/month), my limits comfortably covered a full month of work, and Pro+ gave even more room. After the change, I subscribed to the new Premium plan, and using it the same way I always have (Grok Bot + Cursor), I hit the limit on the very first day.
I reached out to support to understand what happened, and their suggestion was to move to Super or Ultra. For someone coming from Pro, that’s roughly 5x the previous cost to keep the same routine. And since Premium lasted one day for me, it’s hard to know how long a higher tier would last.
I’m sharing this because I enjoy the product and want to keep using it. A few things that, in my view, would help people in a similar situation:
- A middle-tier option or a transition period for users coming from Pro/Pro+
- More clarity on how much each product uses from the usage limit (e.g., Grok Bot)
- Tips on using Grok Bot more efficiently
Curious to hear if others have had a similar experience.
I upgraded from Pro to the new Premium plan today. Shortly after, I used Grok Bot for one short, casual conversation while making breakfast:
- one web lookup for a reasonable price on a product
- a brief chat about two stocks
That was all. No coding, no agents, no long context. When I checked afterwards, 7% of my weekly included usage was already used (resets Oct 21).
At this rate, a few light chats would use up a whole week’s quota. That doesn’t match how the plan was presented, and it feels nothing like the value I had on Pro.
Questions:
- How is Grok Bot usage counted against the weekly included quota? Do web searches or hidden context cost extra?
- Is there a per-request breakdown so I can see what actually used the 7%?
- Could this be connected to the Grok Bot service degradation incident that is active right now? If so, will incorrectly counted usage be refunded?
Screenshot of the usage meter:
At $20 a month, I could justify it because I just use this for light tasks: no coding, social media posting, things like that. At $30 a month, you’ve lost me.
no we need the 20$ as the best plan no 30 $
Cursor would not allow me to sign up before, it rejected my debit cards and bank info through Stripe. I Subscribed to the new bundled X Subscription solely to be able to use Cursor, and after confirming my subscription, Cursor does not recognize the new bundled X Subscription. Cursor tells me subscribe to premium to use the service, which I already have.
The unification interface, both the one from cursor itself and the one from xAI, throw errors and refuse to work.
Cursor support told me on the 12th that they’d have it fixed in a couple hours, or the next morning by the latest. Yesterday morning I reached out for further assistance, and they ignored me. Afterwards, I created a new support ticket, and they closed it and told me to shut up. They have been ignoring me for 2 days now. Very disrespectful.
Support has also insisted multiple times that my old Cursor subscription will still be valid and secure me access to the service while the integration is being fixed, and they have ignored my several protests where I’ve told them I have no prior Cursor subscription and am locked out of the service.
Positively useless customer service.
I think this is likely a bug on their end.
Cursor’s core value has always been model freedom: picking the best model for each task, whether that’s Claude, GPT, Gemini or Grok. Removing all included third-party usage from the new plans undermines exactly that. If Grok is better, let it win on merit, not through pricing that pushes users away from other models. Please keep meaningful third-party allowance in the new plans, and give legacy users a clear, long-term guarantee before any forced migration.
I’m developing software with Cursor with Pro+ plan, and I’m feeling good with Cursor Grok models.
I don’t use or need any other services from X Corp, i.e., Grok Chat, Grok Bot, or X, so please don’t simply bundle them into every plan, force me to pay for and use them.
I suggest you bring back a new range of subscription plans, that contain purely Cursor usage, similar to current Pro, Pro+ and Ultra plans. Or, just leave these plans available again.
If I don’t merge, can I continue to renew the old package indefinitely.
Don’t forget you no longer get access to 3rd party models. So your usage is actually a lot less than they let on.
I am also wondering about this. I have no interest in X, I live in the real world.
We need a clear answer for users that want a distraction free environment and only use Cursor.ai for engneering tasks.
我想继续用20美元的订阅，因为我目前只有支付宝付款，没有其他国际付款手段。能帮忙处理下不
在 2026-10-14 21:08:29，“Colin” [email protected] 写道：
Is it possible to bring back Alipay as a payment method for China?
I was previously on the Pro+ plan ($60/month) and recently switched to the unified Super subscription ($100/month). However, I’m seeing almost the same token usage on both plans.
During checkout, it said I’d get 2.25x more usage than my current limit, but I’m not seeing any difference. Did I miscalculate, or is there something I’m missing?
Also, I never had a Team plan before, but after moving to the unified super plan I now have access to Team features, which I don’t need at all.
Has anyone else had this experience? Do I need to adjust any settings on my end?
Well I took the plunge, and my assessment after 24 hours is, I’m getting twice the usage for a model that takes over 4x the effort to get anything done. 75% of that time has been spent swearing and yelling at my monitor wondering where the I in AI went. I don’t know who came up with the term “vibe” coding, but this is anything but a good vibe.
It seems like Musk might be overstating the credit limits; from what I’ve seen, some people’s assessed limits aren’t nearly as high as what was officially claimed.
I’m currently on a legacy Cursor Ultra annual plan. My dashboard shows two included buckets that reset monthly: “Cursor Models (includes Cursor Grok and Composer)” and “Other Models”. I’m considering switching to The X Subscription Ultra ($2,000/yr). Since switching is permanent, I’d like to clarify two things before I do.
1. What does “2x usage compared to your current plan” actually measure?
The unify pricing page says Ultra gives “2x usage compared to your current plan”, and the docs describe a single shared usage pool that resets weekly. My current plan, however, has a monthly allowance. So:
- Is the 2x comparison against my current monthly “Cursor Models” allowance, or against that allowance converted to a weekly equivalent?
- Put concretely: will each weekly pool give me about 2x my current monthly “Cursor Models” allowance, or about 2x of (monthly allowance ÷ ~4.3)?
- Is there a published absolute number, or any way to see the weekly pool size in dollar terms or as a percentage of my current plan before switching?
- Since Grok Bot, the Grok apps and Cursor all draw from the same weekly pool at different rates, does the “2x” figure assume the whole pool goes to coding in Cursor with Grok models?
2. How is Grok Bot usage billed on the unified plan when the underlying model is a third-party model?
The docs say third-party models in Cursor (e.g. Claude, GPT) are not part of the included usage and are charged at API rates from the first use. As a user, I can’t see or choose which model Grok Bot uses behind the scenes. So:
- On The X Subscription, is all Grok Bot usage always drawn from the shared weekly pool, regardless of which model it runs on internally?
- Or can Grok Bot usage ever be billed separately at third-party API rates (from Additional Usage credits / on-demand spend)?
- If Grok Bot does use third-party models internally, is that usage discounted the way it effectively is today, or charged at full API list price?
Thanks a lot. A clear answer on these two points would help me (and probably many other legacy Ultra users) decide whether to switch.
I only want to use Cruser.
I don’t need X.
I can’t keep up with Elon Musk’s whims.
I only want to use Cruser’s Codex.
Most of us are on Composer 2.5 with a Pro+ subscription.
I don’t need Grok, and I couldn’t care less about X.
I don’t need ad-free scrolling on most screens.
I don’t need the “Preferred @Username” feature.
I don’t need the “Further Improve Reply Priority” feature.
It’s extremely disappointing that they’re adding unnecessary “value-added” features—which are essentially price hikes—while cutting core functionality.
I have absolutely no intention of using X—it’s full of money-grabbing fake videos, staged content, and posts designed to provoke and trap people.
I just want to use Cruser, plain and simple.
I don’t need X—doesn’t anyone else feel the same way??
me too