SuperGrok Pricing: What Each Tier Actually Buys

SuperGrok is not one product. It's five of them, priced from zero to maybe $300 a month, and xAI's own pricing page only confirms the cost of three.

That gap is where most of the confusion lives.

Ten different articles about SuperGrok give you ten different numbers for the top tier. $60. $300. Sometimes no number at all, just a button that says "Get Heavy." The articles all rank fine. They just don't agree with each other, and most of them never opened xAI's actual pricing page before they published.

Here's what's actually there this week: what each tier buys, where the confirmed numbers stop and the guesswork starts, and whether any of it beats the subscription you might already be paying for somewhere else.

Three questions matter more than the marketing copy. What does your card actually get charged. What happens the week you use more than xAI expected. And does paying one vendor for more of one model beat paying less for several.

A fourth question hides under those three: who actually owns the company setting these prices, and did that change recently enough that nobody's updated their review yet. It did. Almost nobody's caught up.

What SuperGrok Costs Right Now, According to xAI

I checked x.ai/pricing directly on October 9, 2026. xAI also runs an entry option that doesn't charge anything, with real-time web and X search, voice mode, and connectors, at far lower limits than either paid tier below it.

Two tiers above that carry a visible dollar figure on the page itself.

That's it. Two confirmed numbers, straight from xAI's pricing page, plus the entry option above them.

Scroll past those cards and the page's own comparison grid lists four more columns: SuperGrok Lite, SuperGrok Heavy, Business, and Enterprise. None of them show a dollar figure in that grid on the page I read.

You click "Get Lite" or "Get Heavy." The price only shows up once you're already in the checkout flow.

That's an odd choice for a company selling a subscription. Our breakdown of Claude's pricing lists all five of Anthropic's tiers with a number attached to every one, on the page you land on, before you touch a card. xAI does that for two tiers out of seven and makes you click to find the rest.

Third-party trackers fill the gap. A pricing explainer puts SuperGrok Heavy at around $300 a month (AI Toolbox). A consulting firm's pricing guide lists SuperGrok Lite at $10 a month, noting the figure was "verified May 19, 2026, per xAI" (TechJack Solutions).

Treat both as likely, not confirmed. Pricing pages move. Before you hand a card number to either tier, do this instead of trusting a blog post:

1. Go to grok.com or x.ai/pricing directly, not a search result.
2. Click the upgrade button for the specific tier you want.
3. Read the number on the actual checkout screen before you enter payment details.
4. Screenshot it, with the date visible, so next month's renewal doesn't surprise you.

Step three is the one people skip. It's tempting to trust the price implied by the upgrade button itself, but that button often just says "Upgrade" with no number attached, and the real charge only confirms on the next screen, sometimes after a regional tax calculation nudges it slightly from what the marketing page implied.

Four steps. Takes two minutes. Saves you from paying whatever a content farm guessed six months ago.

If you're trying to track every AI vendor's pricing shifts as they happen rather than guess at them, that's the whole job of our artificial intelligence coverage: new models, new tiers, and what changed since the last post.

What $30 a Month Actually Buys on the SuperGrok Tier

The entry paid tier gives you four things that matter, according to xAI's own page: the Grok 4.6 model, meaningfully higher rate limits, Expert mode for harder reasoning, and image and video generation through Grok Imagine.

The model bump is real. Grok 4.6 is a current frontier release, built for coding, knowledge work, and STEM, with a 500K token context window and vision input. That's a genuine step up from xAI's no-charge option, which xAI describes only as "generous limits" on an older model line.

Grok's flagship has moved fast. It's climbed through several full version jumps since Grok 4 first launched, and each jump has quietly redefined what a given tier's allowance is actually buying you that month. A $30 subscription bought you less model two releases ago than it buys today, at the same price.

The rate limit bump is the part worth testing before you pay. xAI calls it "higher rate limits across all features," with no number attached. That phrase covers chat messages, DeepSearch queries, and image generations, all drawing from one shared allowance that resets weekly.

Picture a real week. You're running competitive research on three vendors, drafting a long proposal, and generating a dozen product images for a launch post. On the lowest tier, you'll hit a wall partway through that list. On SuperGrok, you probably won't, unless you're doing all three every single day.

Expert mode is the reasoning toggle. It makes the model think longer before answering. That helps on math, multi-step planning, and anything with a wrong answer you can check against a right one.

It also burns through your weekly allowance faster. Leave it on by default and you'll find the ceiling sooner than you expected.

Image and video generation through Grok Imagine is where a lot of SuperGrok subscribers actually spend their allowance, not on chat at all. If your real use case is marketing visuals or short clips, the $30 tier is competing directly with a dedicated image tool. Price it against that, not against a chatbot subscription.

Grok Build, the coding agent bundled into every tier, draws from that same weekly pool once you're on a paid plan. Run it against a real refactor, not a toy script, and you'll see the allowance move fast. Factor that in if coding is your actual use case rather than chat.

SuperGrok Plus at $100: Who Actually Needs the Second Tier

SuperGrok Plus adds one concrete thing and one vague one. The concrete thing is 1080p video generation, up from whatever resolution the base tier caps you at. The vague thing is "significantly higher usage across Chat, Imagine, Voice & Build," plus faster replies and priority access when things are busy.

Do the resolution math first, because it's the only number here you can actually verify.

1080p carries roughly 2.25 times the pixel count of 720p. If your output lands full-screen somewhere, that's the entire argument for paying $70 more a month than the base tier.

Picture an actual job: a 90-second product video for a client who wants it uploaded straight to their website, full width, no cropping. Rendered at 720p, that clip looks soft on a 27-inch monitor, and a client paying for video work will notice before you do. Rendered at 1080p, it holds up. That's the $70 a month earning its keep.

Now picture the opposite job: a 15-second teaser for an Instagram Story. The platform crops it to a vertical frame and compresses it hard before anyone's phone shows it. The gap between 720p and 1080p disappears somewhere in that compression pass. You paid for resolution nobody's eyes will register.

Before you pay the extra $70, answer two things honestly. Does your output get viewed full-screen, uncompressed, by someone who'd actually notice the difference. And have you waited on the base tier more than once this month, for real, not hypothetically.

Two "no" answers means stay on the $30 tier and revisit the question in a quarter.

Priority access is the rest of the pitch. Track it for a few days before you spend the extra $840 a year. If you're hitting a queue only during one predictable hour, like the morning after a product launch when everyone's testing at once, that's a $70-a-month insurance policy for one bad Tuesday. If you can't remember the last time you waited, it's $840 solving a problem you don't have.

Lite and Heavy: the Two Prices Nobody Can Agree On

This is the section every other SuperGrok article skips. Answering it honestly means admitting you can't fully answer it.

SuperGrok Lite sits below the $30 tier, aimed at light users who want basic image and short-clip generation without the full price tag. SuperGrok Heavy sits at the top, aimed at people running continuous, heavy reasoning or multi-agent workloads who treat rate limits as a real cost of doing business.

What matters more than the exact number is this: not one of the review sites I read while researching this post cross-checked their figure against xAI's live page. They cited each other, or cited "xAI" generically. The numbers mostly agreed by coincidence, not because anyone confirmed them that week.

Some didn't even agree with each other. One tracker put Heavy at $60 a month. Most put it at $300. A fifth of what the rest reported is not a rounding error. It's somebody who never clicked through to a real checkout screen.

Run the same gut check on Lite that you'd run on Heavy. If the $10-a-month figure is accurate, that's roughly 33 cents a day for basic image and short-clip generation at a noticeably lower resolution ceiling than the $30 tier. If your use case is a handful of social graphics a week, not daily production work, that's a real gap between what Lite might cost and what SuperGrok charges for capability you'd rarely touch.

If you're seriously considering Heavy at a price that high, don't take any single article's word for it, including this one.

1. Log into grok.com directly.
2. Click through to the Heavy checkout screen specifically.
3. Read the number charging your card, not the number a review quoted.
4. If it's materially different from what you expected, wait a day and decide with a clear head.

The honest test is the same four-step checkout check as the tier above. If Lite's real price matches what trackers report, it's a reasonable bridge between the lowest tier and the full subscription. If it doesn't, you've lost nothing by checking, and you've avoided guessing with real money. A $300-a-month subscription is not the place to trust a number three blog posts removed from the source.

The Shared Weekly Pool, and What "Higher Limits" Really Means

Every paid SuperGrok tier draws from one shared allowance split across chat, image generation, voice, and the coding agent. That's the "pool" xAI's pricing page refers to when it says usage is higher on paid tiers. It resets weekly, not daily, and that single detail changes how you should pace it.

A weekly pool punishes burst usage. Spend Monday generating forty images for a campaign, and you've drawn down the same pool that needs to cover Tuesday through Sunday.

Here's a worked week to see it coming instead of hitting it blind.

1. Monday: competitive research with DeepSearch, roughly 20 to 30 queries. Light draw on the pool.
2. Tuesday: drafting and revising a long proposal in chat. Moderate draw.
3. Wednesday: generating a dozen product images for a launch post. Heavy draw, all at once.
4. Thursday through Saturday: normal chat use, a few DeepSearch checks. Light to moderate draw.
5. Sunday: pool resets, whatever you didn't use doesn't carry over.

None of those five days are unusual for someone running a small brand's content calendar. The pool isn't designed around an edge case. It's designed around exactly that kind of week, which is why Wednesday's image batch is the day most SuperGrok subscribers report hitting a wall.

Burn through the pool by Wednesday afternoon, and Thursday through Sunday you're either waiting, paying for the next tier up, or back on the lowest tier's limits until the reset.

This is the one number worth tracking yourself instead of reading about. Open a plain note on your phone this week. Every time Grok tells you it's rate-limited or under heavy load, log the day and what you were doing.

By Sunday you'll know if you're dealing with a once-a-week edge case or a daily collision. That tells you more about which tier fits than any comparison table will.

If the log comes back nearly empty, the lowest tier was never actually holding you back. Don't upgrade on a hunch.

SuperGrok vs X Premium+: Two Different Purchases

This is the single most common point of confusion in everything written about SuperGrok, and it's confusing on purpose as much as by accident.

SuperGrok is a standalone subscription from xAI, bought at grok.com or x.ai, with no social-platform attachments. X Premium+ is a separate subscription from X, the social platform, priced at roughly $40 a month, that bundles verification and a reduced ad load with a lighter version of Grok (TechJack Solutions).

They're different purchases. Paying for one doesn't give you the other.

Here's where the bundle actually trips people up. Somebody who already pays for X Premium+ for the blue checkmark assumes the Grok access bundled in is equivalent to SuperGrok, because both pages describe it using the same model name. It isn't. The reported rate limits on the bundled version are tighter, and features like DeepSearch have historically rolled out to the standalone tier first.

If you're that person, and you've been hitting walls inside X thinking you already had the premium Grok experience, you don't. You have the lighter version bundled into a social subscription. The $30 standalone tier is a separate purchase, not an upgrade path from what you're already paying for.

Check which one you're actually logged into before you complain about a limit. It takes one glance at the address bar: grok.com or x.ai means SuperGrok, x.com means the bundled version.

If what you want is Grok's full capability and nothing about X the platform, SuperGrok at $30 a month is the direct route and the cheaper one. If you're already paying for X verification and the Grok access bundled in is a bonus you'll use occasionally, X Premium+ covers both in one bill.

Pick based on which product you actually want more of. Don't buy X Premium+ hoping it quietly becomes your AI subscription. Don't buy SuperGrok expecting verification or a lighter ad load to come with it.

What Changes Now That xAI Is Part of SpaceX

Almost nothing written about SuperGrok mentions this. It's the one fact that should change how you think about committing money to it long term.

On February 2, 2026, SpaceX acquired xAI outright in an all-stock deal, combining the two into an entity reported at roughly a $1.25 trillion valuation, with xAI investors swapping their shares for SpaceX stock (Reuters). xAI is no longer an independent company making its own standalone bets on consumer AI pricing.

It's a subsidiary now. Inside a rocket-and-satellite company. With its own enormous capital priorities that have nothing to do with chatbots.

Scale matters here. Before the acquisition, xAI's standalone revenue run rate outside of X's existing subscription business was reported at roughly $500 million a year (Sacra), a fraction of SpaceX's launch and satellite revenue. A consumer AI subscription tier is a rounding error on a balance sheet that size. Pricing decisions for SuperGrok now compete for attention against decisions involving billions of dollars of rocket and satellite infrastructure.

The merger didn't slow xAI's shopping list. Four months later, SpaceX signed a reported $60 billion deal to acquire Anysphere, the company behind the AI coding tool Cursor, explicitly to strengthen Grok Build's position in AI coding (Reuters).

SpaceX didn't do that deal to save money on Grok's consumer tiers. It did it to win a different fight in AI coding. SuperGrok's price sits downstream of decisions made for reasons that have nothing to do with you.

SuperGrok probably won't disappear or get worse tomorrow because of any of this. But the company setting your monthly price now answers to rocket launch schedules and satellite constellations as much as it answers to chatbot subscribers. That's not a reason to panic about next month's charge. It's a reason to treat SuperGrok's long-term roadmap the way you'd treat any product owned by a much larger company with much bigger priorities: useful today, not guaranteed to stay anyone's main focus.

Do the Math Before You Pay

Skip the comparison charts for a minute. Run your own numbers instead.

1. Spend three to five days on xAI's lowest tier doing your normal work. Change nothing else.
2. Log every rate-limit message you hit and what you were doing at the time.
3. If that log stays nearly empty, you don't have a SuperGrok problem. Stop here.
4. If you're hitting it daily, during the exact hours you need Grok most, that's your real signal.
5. Price your own time against the $30 difference. That's roughly a dollar a day.
6. If one rate-limit hit costs you fifteen minutes of waiting, and that happens more than twice a week, the base tier has already paid for itself before the month is out.

Track for five days, not three, if your week includes a weekend push or an irregular schedule. The goal is a representative week, not just a representative Tuesday.

Here's what that log actually looks like in practice. Say you're a solo marketer running three brand accounts. Monday and Tuesday you barely touch Grok. Wednesday you hit a rate limit twice while batch-generating product shots. Thursday and Friday, nothing. That's two hits in a week, both on one day, both predictable.

Two hits a week at fifteen minutes each is half an hour of friction a month. Thirty dollars buys that half hour back, plus everything else the paid tier adds. That's a defensible trade.

Compare that to someone running DeepSearch for client research every morning, hitting limits by Wednesday lunch most weeks, and losing real billable hours waiting for a reset. That's a different number entirely, and it justifies the upgrade faster than any marketing page will.

Only consider Plus or Heavy once you've lived on the tier below it and genuinely hit its ceiling, not hypothetically. Jumping straight to the $100 or $300 tier because an article told you serious users need it is how you end up paying for capacity you never measured needing.

Where SuperGrok's One-Model Bet Costs You

The trade nobody selling you a SuperGrok subscription will spell out is this one. Every dollar of that $30 to $300 buys you more of exactly one model, from exactly one vendor, that just became a subsidiary of a rocket company.

Our head-to-head against ChatGPT found that the real cost of picking one model isn't the subscription price. It's the switching cost you pay the moment that model has an off day on a task it's genuinely weak at, with no second opinion sitting in the same conversation. That post found specific tasks where each model pulls ahead, and specific ones where it trails, which is normal for any single model. The problem isn't that Grok has weak spots. Every model does. It's paying $30 to $300 a month for one model's weak spots with no second opinion in the same window.

None of that is a knock on Grok specifically. The same math applies to paying for ChatGPT alone or Claude alone, without ever checking the answer against anything else. SuperGrok just happens to be the one currently asking the most money for the deepest single-vendor commitment, between its $30 floor and its reported $300 ceiling.

Grok 4.6 and the newer Grok 4.7 are both genuinely strong, and both already sit inside Magai next to the rest of that lineup, in the same thread. Switching models mid-conversation there doesn't reset your context the way opening a second app would.

Inside Magai, that comparison takes under a minute. Type the prompt once, send it to Grok, then resend the identical prompt to a second model in the same thread without losing anything either model already said. No new tab, no copying context between two separate logins, no re-explaining what you're trying to do.

Pick one task you'd actually pay $30 to $300 a month to run through Grok alone. Run it inside Magai's $40-a-month Pro plan instead, where Grok sits beside the others, backed by the same 30-day money-back guarantee that covers every Magai plan.

Compare the two answers side by side before you commit a subscription to finding out the hard way. Start at app.magai.co.