X Subscription Pricing Explained: Grok, Cursor and X Premium Tiers

On October 12, 2026, SpaceXSI (the Musk company that houses Grok and X) announced “The X Subscription”, a single plan family that replaces separate SuperGrok, Cursor Pro and X Premium+ subscriptions. If you pay for any of those, or you are choosing an AI coding tool for a small team, the new structure changes the maths. This guide explains the tiers, what the multipliers mean and who should pick what. Prices are as of October 2026 and come from press coverage of the launch, so confirm them at checkout.

How much does the X Subscription cost?

There are five tiers, from free to $200 a month. Coverage of the announcement lists them as follows.

Tier	Price per month	Usage vs Plus	Best for	Key limitation
Free	$0	Basic allowance	Trying Grok and Cursor’s agents	Small allowance
Plus	$8	1x	Occasional AI users, creators	Entry usage limits not published
Premium	$30	4x	Daily coders and knowledge workers	Third-party models billed separately
Super	$100	15x	Heavy agent and coding users	Same separate billing for outside models
Ultra	$200	50x	Power users running agents all day	Highest cost, personal plan

The company says Premium now covers the benefits of SuperGrok, Cursor Pro and X Premium+, which it says cost about $90 a month together. That is a company claim; your own saving depends on which of those three you actually pay for today.

How do the usage multipliers work?

Paid tiers share one usage pool across Cursor, Grok Build and other Grok products, scaled against Plus. Premium gives four times the Plus allowance, Super 15 times and Ultra 50 times. SpaceXSI has not published how many requests or tokens the Plus tier includes, so the multipliers tell you relative size, not absolute capacity.

Simple arithmetic on the published prices shows the effective cost per “unit” of Plus usage:

• Plus: $8 for 1x, so $8.00 per unit
• Premium: $30 for 4x, so $7.50 per unit
• Super: $100 for 15x, so about $6.67 per unit
• Ultra: $200 for 50x, so $4.00 per unit

Higher tiers are cheaper per unit, but only if you will really use the extra capacity. Most individual professionals will not fill a Super plan.

What happens to third-party models in Cursor?

This is the detail that matters most for developers. Reports say non-Grok models in Cursor, such as those from other AI labs, no longer come with a flat allowance on the new plans. They are billed per request at API list prices, on top of the subscription. One report adds that cloud agents previously available on a $20 Cursor Pro plan now require at least Premium at $30.

The company also claims each paid plan gives at least twice as much coding usage with its own models as the matching Cursor plan. That is a vendor claim, and it only helps if you are happy to work mainly with Grok models.

In practice there are two kinds of developer:

• Grok-first: you will likely get more included usage than before, if the claim holds in your workflow.
• Model-agnostic: if you rely on Claude or GPT models inside Cursor, your bill becomes subscription plus metered API usage, which is harder to forecast.

For a wider view of how coding assistants are priced for teams, see our AI coding assistant pricing guide.

Should I switch from my current subscription?

Probably not by default. The company says existing SuperGrok, Cursor and X Premium subscribers do not have to switch, and anyone who does gets a prorated refund on the old plan. That removes the pressure to decide on day one.

A sensible approach:

1. List what you pay today for SuperGrok, Cursor and X Premium.
2. Note how often you use third-party models in Cursor and what that costs.
3. Compare the total with Premium ($30) first, not Ultra.
4. Switch only if the single plan covers what you use, and keep the refund in mind.

If most of your value comes from X features such as the checkmark, creator tools or fewer ads, Plus at $8 may be enough on its own.

Which tier fits a solo developer or a small team?

For a solo developer who codes daily with Grok, Premium is the natural starting point. For a freelancer who runs long autonomous agent sessions, Super is worth a one-month trial, with usage tracked. Ultra only makes sense if you regularly hit the Super ceiling.

For a small team, remember these are individual subscriptions. Mixing tiers is normal: Premium for people coding every day, Plus for occasional users. Teams in the EU and Latin America should also check how taxes and local pricing appear at checkout, because coverage quotes US dollar list prices. Compare that with the all-purpose assistants in our ChatGPT vs Gemini vs Claude for business comparison if you need general chat and documents rather than coding.

What would a three-person team pay per year?

Take a small product team: two developers who code every day and one founder who mostly uses Grok for research and writing. On the launch prices, a sensible starting mix is two Premium seats and one Plus seat.

Setup	Monthly cost	Yearly cost
2 x Premium + 1 x Plus	$68	$816
3 x Premium	$90	$1,080
2 x Super + 1 x Plus	$208	$2,496

These figures cover the subscriptions only. If the developers keep using Claude or GPT models inside Cursor, add the metered API cost on top, and that is the part to watch. A practical rule is to set a monthly budget for third-party model usage per developer, check it after the first two weeks and adjust before the bill closes. Taxes such as VAT are not included in the list prices above.

The jump from the first row to the third shows why upgrading everyone “just in case” is expensive. Upgrade one seat at a time, and only when someone actually hits the limit.

How should you track usage after you subscribe?

Treat the first month as a measurement period rather than a commitment. Because SpaceXSI has not published the absolute size of the Plus allowance, the only reliable number is the one you observe in your own account.

A simple routine works for most people:

• Week 1: note how quickly you consume the allowance on a normal working day.
• Week 2: check whether any third-party model requests appear on your bill, and how much they cost.
• Week 4: compare the total with what you paid before for SuperGrok, Cursor or X Premium.

If you never come close to the limit, move down a tier. If you hit it most weeks, test the next tier for one month and compare. For teams, put one person in charge of collecting these numbers so decisions are based on usage, not on the tier names.

How we chose what to cover

This guide is based on the announcement as reported by several outlets, on the Cursor pricing page and on public documentation. We did not run side-by-side benchmarks. Where reports disagree or the company has not published a figure, such as the Plus usage limits, we say so instead of guessing. Earlier reporting of a four-tier plan with different prices was superseded by the launch pricing described above.

What are the risks of a bundled plan?

Bundles simplify billing but add dependency. If one plan covers your coding tool, chatbot and social features, a price change or model change affects all three at once. The shared usage pool also means heavy use of one product can reduce what is left for another. Review your usage after the first billing cycle, and keep an export of your key projects and settings so you can move if the economics change.