On October 12, 2026, xAI did something nobody in the AI world saw coming: it rolled Grok, the Cursor coding editor, and X Premium into one subscription bill, with prices from free to $200 a month. The headline claim is that the $30 Premium plan bundles about $90 worth of separate subscriptions. I’ve spent the last few days checking every tier against xAI’s and Cursor’s own pricing pages, and here’s the honest verdict: the math genuinely works — but only for one specific kind of user. If Grok’s models do your work well, this is the best deal in AI coding right now. If your workflow lives on Claude or GPT, the same plan can quietly cost you more than before.
Let me walk through the tiers, the catches, and who should grab this versus who should stay exactly where they are.
What actually changed (and when)
For years, these were three separate products with three separate bills. You could pay for SuperGrok (xAI’s premium chatbot tier), Cursor Pro ($20/month for the AI code editor), and X Premium+ through three different portals, with three different usage meters and three different renewal dates. Then came the ownership change: SpaceX completed its all-stock purchase of Anysphere, Cursor’s parent company, on August 14, 2026, folding xAI into what press calls the SpaceX Super Intelligence group.
The new plan, called The X Subscription, collapses all three into one bill, one account, and — this is the part that matters — one shared pool of AI usage that both Grok and Cursor draw from. xAI’s announcement post is dated October 2, and the new pricing went live on x.ai/pricing and cursor.com/pricing around October 12–13, 2026. The rollout is gradual: some users still see the old plans in their billing settings, and xAI says existing subscribers can keep what they have — switching is optional for now. The launch was widely covered, including TechCrunch’s reporting on the bundle, which confirms existing X Premium subscribers can switch from their subscription settings while the option rolls out.
One thing to know before we get into numbers: for new customers, the old Cursor plans are closed to new sign-ups. Cursor’s pricing page now sells The X Subscription instead of Pro ($20), Pro+ ($60), and the old standalone Ultra ($200). If you’re already on a legacy plan, nothing changes unless you choose to switch.
The five tiers, side by side
Here are the tiers as xAI lists them. Prices are monthly, in US dollars, before tax. The annual options (Plus $80/year, Premium $300, Super $1,000, Ultra $2,000) work out to roughly two months free.
| Tier | Price | Included usage | What you actually get |
|---|---|---|---|
| Free | $0 | Limited | Limited Grok chat and agent requests, a starter allowance for Cursor’s coding agents, standard X |
| Plus | $8/mo | Base pool | Grok Chat, Voice, Imagine and Build, Grok API access, Cursor (Grok models only), X checkmark, fewer ads, creator tools |
| Premium | $30/mo | 4x Plus | Everything in Plus, plus Grok Bot, the newest Grok models, faster chat, AI images and video, cloud agents and automations, code hosting, AI code reviews, ad-free X on most surfaces, preferred handle, reply boost |
| Super | $100/mo | 15x Plus | Everything in Premium, plus priority model access at peak times |
| Ultra | $200/mo | 50x Plus | Everything in Super, plus earliest access to new products and models |
There are also Teams seats at the same $30/$100/$200 per-seat prices as Premium, Super, and Ultra, with SSO, admin controls, centralized billing, and usage analytics — plus custom Enterprise pricing. If you run a small dev shop, the Teams option is the first way to buy Cursor for your whole team without the old $40-or-$120-per-seat legacy Teams math.
Two structural facts shape everything below. First, the usage allowance refreshes weekly, not monthly. After you burn through it, you wait for the refresh or buy top-ups. Second, xAI publishes usage only as multiples of Plus — it never tells you the absolute dollar size of any tier’s pool. That opacity matters when we get to the “should I switch” math.
The “$30 buys back $90” claim: is it real?
xAI’s pitch for the Premium tier is blunt: “The benefits of SuperGrok, Cursor Pro, and X Premium+, previously $90/month combined.” That is a company claim, and it’s worth testing instead of repeating.
Here’s what checks out. Cursor Pro was genuinely $20/month for new buyers — nobody disputes that. The combined-statement math is the part xAI supplies, and every independent pricing breakdown I’ve found (from the pricing analysts who read x.ai/pricing line by line in mid-October) lands on the same conclusion: if you were already paying for all three products separately, $30 is a steep cut. You’d need to be running three separate bills totaling roughly $90 a month for the full claim to be true in your case.
But — and this is the catch that one analysis site, winzheng.com, nailed in its review — the $90 figure describes the maximum-stack buyer. Their verdict, in plain terms: “For users who need an AI assistant, coding tools, and enhanced social platform features at the same time, the book value of $30 for $90 is real, and migrating is worthwhile.” Then comes the condition, which I’ll put in bold because it’s the whole story: you have to be willing to replace Claude or GPT with Grok as your primary coding model.
If you weren’t paying for all three before, you’re not saving $90. You’re deciding whether Grok + Cursor + X Premium, as a bundle, is worth $30 a month compared to what you’d actually buy otherwise. That’s a different question, and it’s the one the rest of this article answers.
The real trade: one pool, Grok-first
Here’s where the deal turns. Under the new plans, only Grok models draw from the included usage pool. On the $8 Plus tier, Cursor runs Grok models only — no third-party models, no cloud agents, no Bugbot code reviews. On Premium and above, Claude, GPT, and Gemini are still available inside Cursor, but they’re billed separately at API list price through prepaid top-ups. They are completely outside the included pool from the very first request.
xAI is explicit about the upside side of this trade: each plan includes “at least twice as much coding usage with our models as the comparable Cursor plan.” So if Grok does your coding work, you’re getting a genuinely bigger tank of usage than the old Cursor plans offered.
If your work runs on Claude or GPT, though, the bundle economics invert. Consider the developer on legacy Cursor Pro at $20/month who codes with Claude: their included usage covered third-party models. On the new $30 Premium plan, that same developer’s Claude usage bills at API list price on top of the $30. The $30 floor becomes a floor, not a ceiling — your actual bill is $30 plus whatever your model usage costs.
This is the part where I want to be blunt, because several developer commentators have already said it well. On Cursor’s own announcement forum, a long-time Cursor Ultra subscriber wrote: “Cursor earned its place next to Codex and Claude Code not because of its own model, but because it was ‘one editor where you can use the best models.’ … The new plans change that. The monthly fee only buys Grok usage.” Their conclusion — “Users aren’t fooled; they’ll vote with their feet” — is dramatic, but it captures a real sentiment: Cursor’s core value was model neutrality, and the new plans trade that neutrality for Grok volume.
(One forum reply is a small, self-selected sample — developer forums skew negative. But the underlying fact it points at, the pricing structure, is verified on Cursor’s own pages.)
Tier-by-tier: who each plan is actually for
Free ($0) — A taste. Limited Grok chat, limited agent requests, a starter allowance for Cursor’s coding agents, standard X. xAI is clearly hoping you discover Grok and upgrade. If you just want to check whether Grok can handle your workflow before paying, start here and run it against a real task.
Plus ($8/month) — This is the plan xAI recommends for “social creators with limited AI usage,” and the price point is not a coincidence: it matches ChatGPT Go at $8. You get the X checkmark, fewer ads, creator tools, Grok Chat/Voice/Imagine/Build, and Cursor with Grok models only. If you live on X and want a little AI plus light coding with Grok, it’s genuinely well priced. If you code seriously, the missing pieces (cloud agents, third-party models, Bugbot) will push you up a tier fast.
Premium ($30/month) — The plan xAI markets as “most popular,” and the one where the $90-combined claim lives. Four times the Plus usage, Grok Bot, the newest Grok models, video generation, cloud agents and automations, code hosting, AI code reviews, third-party models at API list price, and ad-free X. This is the value pick for daily AI coders who are happy with Grok models. The weekly refresh cadence matters here: one allowance per week is friendlier than monthly for burst-style coding work.
Super ($100/month) — Fifteen times the Plus usage, plus priority model access at peak times. This is for people running agents all day. xAI’s own recommendation: daily agent users. If you spend your workday with cloud agents running refactors in the background, this is your tier. Everyone else should test Premium first.
Ultra ($200/month) — Fifty times the Plus usage, plus earliest access to new products and models. xAI recommends this for power users, and the price mirrors the old Cursor Ultra’s $200 and ChatGPT’s $200 Pro tier. Nobody should start here; graduate into it if you actually exhaust Super.
A quick note on the annual math, since it changes the story a bit: at $300/year, Premium costs you two months’ free equivalent. If you know you’re in for the year, the annual plan takes the effective monthly cost to $25 — which is only $5 more than the old Cursor Pro you might have been paying anyway.
The comparison ladder: xAI vs OpenAI vs everyone else
Nobody buys AI tools in a vacuum, so here’s how the X Subscription stacks up against the parallel ladders — prices verified against current pricing pages:
- $8 tier: X Subscription Plus ($8) vs ChatGPT Go ($8, Luna model only). Plus gives you coding (Grok models in Cursor) and the X checkmark; Go gives you OpenAI’s models in a chat interface. If you code at all, Plus wins. If you only chat, it’s a model-quality question: Grok vs GPT-6 Luna.
- $20 tier: This used to be the Cursor Pro and ChatGPT Plus price point. Cursor’s new Premium is $30 — meaning xAI’s entry point for serious coding went up $10, while ChatGPT Plus stays $20. ChatGPT Plus ($20) includes Codex coding with per-five-hour-window usage. Cursor Premium ($30) buys you the editor, cloud agents, and weekly Grok allowance. If you’re a heavy AI coding user, several analysts note the most common real-world answer: people run both, roughly $50 a month, splitting by where the code lives.
- $100 tier: X Subscription Super ($100) vs ChatGPT Pro $100 (5x usage tier) vs Claude Max $100. All three are “heavy daily use” plans. The X bundle is the only one that wraps a social platform subscription into the deal.
- $200 tier: X Subscription Ultra ($200) vs ChatGPT Pro $200 (now 10x the Plus allowance for new subscribers, down from 20x) vs Claude Max $200. Same pattern: xAI’s differentiator is the bundle and the Grok-model volume; the others are model-quality plays.
- Budget coding-only: GitHub Copilot Pro is still $10/month — worth remembering if you want a code assistant without any social-platform or chatbot bundle attached.
The pattern is clear: xAI’s ladder now parallels OpenAI’s ladder almost price-point for price-point ($8, $30/$20, $100, $200). The industry really is consolidating around these five rungs. The X Subscription’s unique selling point isn’t any single rung — it’s that one bill covers a chatbot, a coding editor, and a social platform.
What real users are actually saying
The early reaction has been louder than the launch. A YouTube breakdown titled “Cursor is now part of the X Subscription. What did they just do?” sums up the skeptical case in one line: “If Grok does your work, the bundle is a good deal. If it doesn’t, you’re paying $30 for a checkmark and buying your real models at full price.”
On the developer side, the winzheng.com analysis frames this as something bigger than pricing: “This is precisely how platform bundling strategies work in their most classic form — not forcing you to switch, but making ‘not switching’ more expensive.” Their prediction: the outcome depends on whether Grok’s model capabilities catch up to developers’ real needs fast enough.
My take after reading all of it: the loudest complaints come from exactly the users the new plans serve worst — multi-model developers who chose Cursor precisely because it wasn’t tied to one vendor. xAI is betting that Grok’s coding quality plus bundled volume wins more users than neutrality loses. It’s a real gamble, and the honest answer to “is it worth it” depends entirely on which side of the Grok question you sit on.
Who should grab it, and who should stay put
Grab Premium ($30) if: you code daily in Cursor, Grok’s models handle your work (test this on a real codebase first), and you already pay for X Premium or wanted SuperGrok. This is the group where “$30 replaces ~$90” is actually true, and the doubled coding usage is real value.
Grab Plus ($8) if: you live on X, want the checkmark and fewer ads, and like the idea of a little Grok coding on the side. It’s the cheapest on-ramp to the bundle.
Stay on your legacy plan if: you’re on Cursor Pro ($20), Pro+ ($60), or Ultra ($200) and your workflow runs on Claude, GPT, or Gemini inside Cursor. Your included third-party usage dies the moment you switch, and one report (StoryBoard18, citing xAI’s documentation) says switching is permanent — so this is a one-way door. Don’t walk through it to save $0.
Test before you commit if: you’re curious but unsure. The practical move several developers recommend: keep your main setup on the legacy plan, and test the new tier on a second account for a couple of weekly refresh cycles. Watch the usage pool, watch the top-up bill for third-party models, and compare the diffs Grok produces on your actual tickets against what your current model does.
Skip the bundle entirely if: you don’t use X and don’t want to, you want a neutral editor with included multi-model usage (the old Cursor value prop), or you’re a light user better served by Free or Plus.
The 10-minute decision framework
Here’s the whole decision compressed into one flow:
- Do you already pay for X Premium or SuperGrok? If yes, price the Premium bundle against your current bills. You might genuinely save.
- Run Grok against your real work. Not a benchmark, not a demo — your codebase, your tickets. If the output quality holds, the bundle’s math is in your favor.
- Audit your model mix. What share of your Cursor usage is Claude, GPT, or Gemini? Those hours move to API list price under the new plans. Price one month of that at list rates before switching.
- Check the weekly rhythm. The allowance refreshes weekly. If your work comes in monthly bursts (sprint weeks followed by quiet weeks), weekly refresh actually suits you. If you do marathon weekends, watch the cap.
- Set a spending cap before you test. Third-party models bill through top-ups — set caps in billing so a heavy Claude day doesn’t surprise you.
And one practical note: prices, usage multiples, and migration terms are all changing fast right now. xAI is still rolling this out gradually, and the details on billing settings pages are ahead of the marketing pages. Check x.ai/pricing and cursor.com/pricing on the day you decide — don’t trust a price list from even two weeks ago, including this one.
The bigger picture: the bundle wars have started
Step back from the price table and the strategy is obvious. xAI now sells the model (Grok), the app (Cursor), and the distribution channel (X) as one subscription — and it’s not alone. OpenAI has its own ladder from Go ($8) to Pro ($500), Meta has Muse Wonder, Google has Gemini 4 Argon. As one commentator put it, what’s being contested in 2026 is no longer the strength of any single feature but the lock-in power of the whole ecosystem.
For users, this means subscription math is now portfolio math. The question isn’t “is Grok good” or “is Cursor good” — it’s “which bundle of model + editor + perks covers the most of what I actually use, for the least money, with the least lock-in.” xAI’s answer is the most aggressive bundle yet. Whether it’s your answer depends on one question only: does Grok do your work? Run the test. Then do the math.
For more on the coding-tool landscape, our coverage of Cursor’s cloud features and OpenAI’s GPT-6 and Codex tiers lays out the rest of the playing field, and our Claude Opus 5.5 pricing breakdown is worth a read if you’re one of the Claude-first developers weighing the switch.
Frequently Asked Questions
What is the X Subscription?
The X Subscription is xAI’s unified plan, launched around October 12–13, 2026, that bundles Grok (chat, voice, Imagine, Build, API, and Grok Bot), the Cursor AI coding editor, and X Premium perks (checkmark, creator tools, fewer ads) into a single bill with one shared AI usage pool. It replaces standalone Cursor plans for new customers; existing subscribers keep their current plans for now.
How much does the X Subscription cost?
Five tiers: Free ($0, limited), Plus at $8/month, Premium at $30/month, Super at $100/month, and Ultra at $200/month. Annual billing runs $80, $300, $1,000, and $2,000 respectively — roughly two months free. Usage scales as a base pool on Plus, then about 4x, 15x, and 50x on Premium, Super, and Ultra, and refreshes weekly.
Does the X Subscription replace Cursor Pro?
For new customers, yes — Cursor’s pricing page now sells The X Subscription instead of Pro ($20), Pro+ ($60), and the old standalone plans. For existing Cursor subscribers, xAI says nothing changes: legacy plans continue and switching is optional, though the migration option rolls out gradually in billing settings.
Are Claude, GPT, and Gemini included in the X Subscription?
No — and this is the most important fact on the page. On the Plus tier, Cursor runs Grok models only. On Premium and above, third-party models remain available but are billed separately at API list price through prepaid top-ups, completely outside the included usage pool. Only Grok-model usage is covered by the subscription. If your workflow runs on Claude or GPT, switching can raise your total bill.
Is the $30 plan really worth $90?
That’s xAI’s claim — that Premium bundles SuperGrok, Cursor Pro, and X Premium+, previously about $90/month combined. It’s real value if you were actually paying for all three. If you weren’t, compare the $30 bundle against what you’d genuinely buy otherwise (for example, ChatGPT Plus at $20 plus a code editor), and run Grok against your actual work before deciding.
Can I switch back if the new plan doesn’t work for me?
xAI says switching is optional for existing subscribers, and the migration rolls out gradually. But at least one report (StoryBoard18, citing xAI’s documentation) says the switch is permanent — so treat it as a one-way door until xAI states otherwise. Test the new tier on a second account before moving your main billing.