Grok Pricing Now Runs $8 to $200 in 5 Tiers: The AI Price Ladder Compared

Grok pricing now runs from $8 to $200 a month in five tiers, and the top tier buys 50 times the usage of the bottom paid tier for 25 times the price. xAI announced the restructure on 12 October 2026 under the name The X Subscription. That makes Grok the only one of the four big consumer AI ladders that gets sharply cheaper per unit as you climb.

The short answer

Grok costs $0, $8, $30, $100 or $200 a month in the US after the October 2026 restructure. Plus at $8 adds more Grok usage and X perks. Premium at $30 adds the SuperGrok, Cursor Pro and X Premium+ bundle at 4x Plus usage. Super and Ultra add no features, only 15x and 50x the usage.

Grok pricing after 12 October: five tiers, one bundle

xAI did not just reprice Grok. It merged three subscriptions into one ladder. The X Subscription covers Grok, the Grok Bot agent product, Cursor's coding agents and X Premium, according to xAI's announcement dated 12 October 2026. The company said it would reach every market over the following days.

The bundling is the real change. Before October, a buyer who wanted Grok, Cursor and the X perks paid three vendors and three bills. xAI says those three products cost $90 a month combined at the Premium level, and Premium now costs $30.

That framing deserves some care. A $60 saving only exists for someone who would have bought all three. A buyer who only wants Grok sees a different ladder, and that buyer is the one this comparison is built for.

What each Grok tier buys

The free tier gives daily Grok access, a starter allowance for Cursor's coding agents and standard X. Plus, at $8, adds more AI usage in Grok and Cursor, plus a checkmark, creator tools and fewer ads on X. xAI labels Plus usage as the base unit for the whole ladder.

Premium, at $30, is the first tier with the full bundle and 4x Plus usage. Teslarati reported on 13 October that Grok Bot is included from Premium upward and not on Plus. xAI's own page says third-party models in Cursor stay available on Premium and above, billed at API list prices through top-ups.

Super, at $100, and Ultra, at $200, are described in xAI's table only as "All of the above features." The difference is usage: 15x Plus for Super and 50x Plus for Ultra. Above Premium, you are buying allowance and nothing else.

What happens to SuperGrok and X Premium subscribers

Nothing changes unless you choose it. xAI says existing SuperGrok, Cursor and X Premium subscribers can keep their current plan, price and features, or switch for a prorated refund on the old plan. Teslarati reported that switching is permanent once accounts are unified, and that the old plans are closed to new sign-ups.

We would not switch on day one. A legacy plan that cannot be rejoined is an option with value, and you give it up the moment you move. Check your actual Grok usage first, then decide whether a rung on the new ladder beats what you already pay.

The four consumer AI price ladders side by side

Here is every consumer tier at the four labs, with the usage language each vendor uses and the main thing each price step buys. Prices are US monthly list prices. The last column says where the figure comes from and when it was checked.

Free tiers are omitted. Google's feature and usage columns come from its own plans page, which served an India locale during research, so Google's US prices rest on press coverage. Google AI Plus is listed at $7.99 at the May 2026 relaunch and $4.99 in June reporting; we use the later figure and flag it below.

Grok pricing per unit of usage, inside each ladder

Every lab publishes its tiers as multiples of another tier. That lets you compute one useful number: what each step costs per unit of the ladder's own base allowance. The unit differs by lab, so read this within a ladder, never across ladders.

The working is short. For Grok, divide each price by its multiple of Plus: $8 over 1, $30 over 4, $100 over 15 and $200 over 50. For Claude, divide by the multiple of Pro, using the weekly allowance. For Google, divide each tier above Pro by its multiple of Pro.

Google Plus is computed as half of Pro, because Google rates Plus at 2x and Pro at 4x the usage of its non-AI plans. At the earlier $7.99 price, the same figure would be $15.98. ChatGPT's multiples changed on 29 October and OpenAI does not publish the new numbers.

Three different philosophies sit inside one table. xAI discounts every step, gently at first and then steeply at Ultra. Anthropic charges the same rate at every step, so a bigger plan is simply more of the same. Google holds the rate flat through Ultra 5x, then halves it at Ultra 20x.

We think Anthropic's flat ladder is the most honest of the three. It tells you that heavy use is not subsidised, and it makes the choice purely about how much you need. xAI's curve says the opposite: the $200 buyer is the customer it most wants, and it will pay in allowance to get them.

Claude's top rung is not identical to its middle rung, though. The pricing page says Ultra has no 5-hour limit for verified accounts, while Max is capped per 5 hours and per week. The weekly rate is flat. The freedom to spend it in one sitting is what the extra $100 buys.

Grok pricing against ChatGPT, Claude and Gemini at each price step

Within a single price band, the four labs sell different things.

The sub-$10 rung: Grok Plus, ChatGPT Go and Google AI Plus

Three labs now sell a paid tier under $10. Grok Plus and ChatGPT Go both cost $8. Notebookcheck reported in January 2026 that Go is ad-supported in the US, while Grok Plus reduces ads on X rather than adding them to Grok.

Google AI Plus was $7.99 at the May relaunch and SlashGear reported a cut to $4.99 on 11 June 2026, with storage doubled to 400 GB. Google's entry tier is as much a storage plan as an AI plan. Anthropic sells nothing below Claude Pro at $20, which we read as a deliberate choice to keep the bottom of its ladder for people who will pay for real work.

The $20 to $30 rung: where Grok stands apart

ChatGPT Plus and Claude Pro cost $20, and Google AI Pro costs $19.99. Grok's middle tier costs $30. That is the one rung where xAI breaks the industry's price point, and the bundle is the reason.

Premium carries Cursor Pro and X Premium+ alongside SuperGrok. If you would use all three, $30 is cheap. If you only want a chat assistant, you are paying $10 more than the rivals' middle tier for products you may never open. For a Grok-only buyer, we would start on Plus at $8 and move up only when the limit bites.

The $100 rung: 5x is the industry default, except at xAI

Claude Max at $100 carries 5x Pro usage. Google AI Ultra 5x at $99.99 carries 5x Pro. OpenAI launched Pro 100 in April 2026 at 5x Plus. Three labs independently settled on the same ratio: five times the $20 plan for five times the money.

Grok Super breaks that pattern. At $100 it carries 15x Plus, which is 3.75 times Premium's allowance for 3.33 times Premium's price. Measured against its own $30 tier, Super is a mild discount rather than a straight multiple.

The $200 rung and above: the volume discount lives here

Every lab sells a $200 tier. Grok Ultra carries 50x Plus. Google Ultra 20x carries 20x Pro. Claude Ultra carries 10x Pro a week with no 5-hour cap. OpenAI no longer publishes Pro 200's multiple.

Only OpenAI goes higher. Trending Topics reported on 1 October 2026 that Pro 500 launched at $500 with 25x Plus usage and an Ultrafast mode. Then on 29 October, BleepingComputer reported that OpenAI doubled usage limits on all Pro tiers and cancelled a planned cut to Pro 200. The full sequence is in the dated timeline of ChatGPT plan changes.

Five pricing patterns now common across AI subscription tiers

Set the four ladders next to each other and the same design choices repeat. Some are shared by all four labs, and some by three.

1. The same four price points. $8 or less, about $20, $100 and $200. Every lab sells at least three of them. The only departures are Grok's $30 middle tier and ChatGPT's $500 top tier.

2. Usage sold as a multiple, never a count. xAI calls Plus "Base." Anthropic says Max is "5x Pro usage." Google measures Plus and Pro against its non-AI plans. BleepingComputer reported that OpenAI "won't provide exact numerical usage limits for each tier." Not one of the 16 paid tiers in our table carries an absolute number.

3. Non-AI goods bundled into the price. Google adds storage and YouTube. xAI adds an X checkmark, Cursor and fewer ads. Anthropic now adds $100 and $200 a month in API credits to Max and Ultra. Bundling moves the comparison away from price per message, which is the comparison every vendor would rather avoid.

4. Coding agents justify the top rungs. Cursor sits inside The X Subscription. Claude Code is included from Pro. OpenAI's Pro tiers are measured in Work and Codex usage. Heavy coding sessions are what burns allowance, so the $100 and $200 rungs are priced for developers who happen to buy consumer plans.

5. Grandfathering as a retention tool. xAI lets SuperGrok subscribers keep their plan. OpenAI cancelled its Pro 200 cut after feedback. Price stability with allowance changes is now the normal shape of a repricing, a pattern covered in the analysis of AI shrinkflation in usage limits.

Why the ladders converge on $20, $100 and $200

The obvious explanation is competition: once one lab prices at $20, a rival at $25 has to explain the extra $5. The less obvious one is psychology. Round prices at clean multiples make the next step easy to justify, because "5x the usage for 5x the price" sounds fair without arithmetic. xAI's $30 rung is the exception that proves the rule, since it needs a $90 comparison to make sense. The trade-off between flat bundles and metered pricing is argued in more depth in the piece on AI bundling against metering.

What the Grok bundle means if you buy AI subscriptions for a small team

This post is written for a founder or director who signs off on AI subscriptions for a handful of people. That reader converts on unit economics, and the ladder comparison is a unit economics question.

xAI says Team users get Premium, Super and Ultra at the same per-seat prices as individual plans. That makes Grok one of the few ladders where a team can mix rungs by person without moving to an enterprise contract. One engineer on Ultra and three colleagues on Premium is a coherent purchase.

The coding usage claim matters for that engineer. xAI says each plan includes at least twice as much coding usage with its own models as the comparable Cursor plan. That is a vendor claim with no published baseline, so treat it as a promise to test rather than a figure to budget on.

The risk is concentration. Putting chat, coding and social into one bill means a single repricing hits three tools at once. We would keep at least one of the three on a separate vendor until The X Subscription has gone through one renewal cycle. If your team is weighing whether to pay for the top rung at all, the comparison of Claude's plans against its API rates shows where a metered API beats a flat plan.

Where this comparison is weakest

No archived captures, and two prices rest on press reports

The brief asked for archived pricing pages. The web archive was not reachable from our research tools, and OpenAI's pricing page refused automated access. xAI and Anthropic rows come from vendor pages read on 31 October 2026. ChatGPT and Google US prices come from dated press coverage. Google AI Plus appears at $7.99 in May coverage and $4.99 in June coverage, and we could not confirm the US price on Google's own page, which served an India locale.

Multiples do not travel between labs

A Grok Plus allowance, a Claude Pro allowance and a Google Pro allowance are different quantities. They reset on different windows and are measured in different ways: Google moved to a compute-used model in May, per The Decoder, and Anthropic counts per 5 hours and per week. Our unit-price table is only valid inside one ladder. Anyone telling you Grok Ultra is "cheaper than Claude Ultra" from these numbers is comparing units nobody has defined. The case for vendors publishing absolute limits is made in the proposed standard for rate-limit transparency.

There is a competing view worth stating. Multiples may be the honest choice, because real cost per message depends on model, length and tool use. A fixed message count would mislead in a different way. We still think a published floor, such as a minimum number of standard messages per week, would serve buyers better than no number at all.

Frequently asked questions

Where to start: price your own usage before you pick a rung

Before anyone on your team switches to The X Subscription, pull one month of usage for each person on each AI tool. Note how often they hit a limit, and on which tool. A person who never hits the Plus limit does not need Premium, whatever the bundle is worth on paper.

Then save today's pricing pages with the date, for every vendor you pay. This post could not use archived captures because the archive was out of reach. Your own dated copy is the only record you control when a vendor changes an allowance and keeps the price.

Related reading