The X Subscription Changes What Cursor Users Pay For

The X Subscription puts Cursor, Grok and X Premium on one bill. Announced on October 12, the new plan also changes a smaller line item with potentially larger consequences for developers: the usage included in its shared allowance is for Grok models. Cursor users who choose eligible third-party models on the new plans pay for those requests separately.

That makes this more than a tidy-up of three subscription pages. Cursor is a coding tool; Grok is an AI assistant; X supplies social-network benefits. The bundle can be attractive to someone who uses all three, particularly if Grok handles much of their coding work. A developer who mainly opens Cursor to use another company’s model has a different calculation. Both customers can buy the same plan and get very different value from the included allowance.

The announced tiers run from a free option with limited access to Grok and Cursor’s coding agents to paid plans at $8 a month for Plus, $30 for Premium, $100 for Super and $200 for Ultra. Plus adds more AI usage and X benefits such as a checkmark and fewer ads. Premium adds benefits the company associates with its former SuperGrok, Cursor Pro and X Premium+ offerings. The higher tiers increase the included usage allowance. For new customers, the company says The X Subscription replaces the previous standalone plans; existing subscribers can keep their current plans for now.

The list of products is broad. The meter underneath it is narrower.

Follow the Allowance

Paid customers get one pool of Grok-model usage across Cursor, the Grok apps and Grok on X. Cursor says that allowance refreshes weekly. A request made with Grok in Cursor and a request made with Grok in an eligible Grok app draw on the same pool. For someone moving between those products, unused allowance in one can be put to work in another.

Cursor also says each new tier includes at least twice as much coding usage with Grok models as its comparable former Cursor plan. It describes the company’s comparison of included Grok usage, rather than a promise that every developer’s monthly bill will fall. How much a person benefits depends on which models they use, how often they use them and whether the X and Grok features replace services they already pay for.

On Premium and higher tiers, third-party models remain accessible in Cursor, but their use is billed separately at API list prices through top-ups. Subscribers can set spending caps. Someone choosing Claude for a coding task, for example, should account for that charge alongside the subscription rather than treating the shared Grok pool as credit toward it. Plus costs less at the outset, but the announced third-party-model option begins at Premium. The cheapest tier on the price chart may therefore be the wrong starting point for a developer whose workflow depends on another provider’s model.

This arrangement has a defensible upside. A subscriber who wants Grok across chat and coding no longer has to divide that included usage between separate product allowances. Separate billing also lets a Premium subscriber use another model without giving up Cursor altogether, while a spending cap offers a way to constrain those charges. The tradeoff is that flexibility in model choice and flexibility in what the subscription pays for have become different things. Cursor still offers the first on eligible tiers; the included pool favors Grok.

The bundle’s non-coding benefits sharpen that tradeoff. Engadget’s breakdown of the tiers lists creator tools and fewer X ads with Plus, while Premium adds Cursor features including code hosting, cloud agents and automations. A person who uses X’s creator tools and Cursor’s agents may find several useful items on the same bill. Someone who wants coding access alone should price the developer features and model requests they actually use. A checkmark cannot pay an API top-up.

What Changed for an Existing Cursor Customer

Cursor’s earlier standalone plans provide a useful baseline. Its former $20-a-month Pro plan included cloud agents and Origin, its GitHub alternative launched in August; both features sit at the $30 Premium level of The X Subscription. For a Pro subscriber who uses those features, comparing the old monthly price with the new Plus price would miss the tier change. Comparing Pro with Premium captures access to those tools, though the model-usage terms still need their own comparison.

The transition is optional today. Cursor says existing subscribers keep their included third-party usage until they switch, and the company is rolling out the option to switch through account billing settings. The X Subscription announcement says existing customers can retain their present price and features; it also says remaining customers will eventually be moved to new plans, with advance notice and no timeline yet. A current subscriber need not interpret the new price chart as an immediate change to an existing bill. Switching, however, changes which model requests the new shared allowance covers.

That history explains why two statements about the launch can both be accurate. The new plans offer more included Grok coding usage than comparable old Cursor plans, by the companies’ account. An existing customer can still lose the former plan’s included third-party usage by choosing the new one. The first claim measures how much of one model family is included; the second concerns what happens when a developer selects a different family. The product retains model choice, but the subscription now puts a price boundary around it.

There is another way to package an AI subscription for developers. In September, OpenAI expanded Sign in with ChatGPT so paying subscribers could use their existing AI allowances in selected third-party developer tools, including Devin, Amp and Warp. The resemblance is that an allowance can follow a subscriber into a coding workflow. The routes differ: The X Subscription carries Grok usage among products in its bundle, including Cursor, while OpenAI’s arrangement extends its subscribers’ allowances into participating outside tools. Those descriptions alone do not establish comparable prices, usage limits or tool capabilities.

In each approach, the model provider has a reason to make its own allowance convenient wherever developers work. For users, convenience is useful only if the included model and the place it can be used match the task. A shared pool across three products is valuable to a person who wants to spend it across those products; it offers little relief for requests billed outside that pool.

The practical comparison starts with a developer’s last month of work, not a promotional multiplier: which models handled the requests, which Cursor features were essential, and which X or Grok benefits would actually be used? The X Subscription has an answer for each item. Whether those answers add up to a better deal depends on the work going through the meter.