Grok pricing: new X Subscription plans, SuperGrok and API costs
If you're new to Grok, it now lives inside The X Subscription, with US list prices starting at $8 a month for Plus and $30 for Premium. If you already have SuperGrok or a legacy X Premium plan, you can keep it and its pricing without interruption. Before you compare prices or sign up for anything else, take a minute to check which billing system your account is on. It can save you from paying twice.
Start with the account you actually have
Ask two people "how much does Grok cost?" and you may well get two different answers, and both could be right. xAI recently restructured its consumer and developer billing, so the answer depends on when and how someone signed up. According to the official account guide, all new individual customers now enroll through The X Subscription. Existing Cursor, SuperGrok, and legacy X Premium subscriptions stay active unless you choose to upgrade or cancel them.
The practical catch: switching billing systems combines selected services and changes your renewal terms. Before you migrate an account or type in a new credit card, open your current account settings and check the cancellation terms, included allowances, and billing currency. It takes a couple of minutes and is worth it.
The X Subscription plans: Plus, Premium and annual billing
The X Subscription is a tiered lineup that covers everyday users, power researchers, and software engineers. Here's how it breaks down:
| Plan Tier | US Monthly Price | Annual Billing Option | Included Tooling & Allowances |
|---|---|---|---|
| Plus | $8 / month | Discounted annual option | Basic Grok access, standard social features, and entry-level generation |
| Premium | $30 / month | $300 / year | Expanded Grok reasoning pool, Grok Bot integration, and priority compute |
| Super & Ultra | Higher custom tiers | Enterprise annual agreements | Maximum concurrency, large team workspaces, and priority queue routing |
If you run high-throughput automation or team-wide development seats, account admins can set monthly usage caps or enterprise billing commitments that scale past $1,000 or $2,000 per month. One thing to keep in mind: promotional checkout discounts only apply to the first billing period. When you're sizing up an ongoing project, judge it by the standard renewal invoice, not the intro price.
What the shared monthly allowance actually covers
The X Subscription includes an included resource pool for eligible first-party usage across official Grok web, mobile, and Cursor environments. Higher tiers (Premium, Super, and Ultra) get progressively larger pools than Plus. The allowance resets on a weekly cadence, so it isn't an unmetered, unlimited-prompt promise. Different activities also draw the pool down at different rates. Generating high-resolution images, for example, uses it differently than running deep code analysis.
Here's the part that trips people up: Third-party models in developer environments like Cursor (such as Claude or GPT models) draw from paid Additional Usage balances from the very first request. Grok Bot is bundled with Premium and higher tiers and uses that same primary seat allocation, so there's no need to budget a separate phantom fee for Bot access.
Marketing materials like to highlight developer API connectivity, but a consumer subscription doesn't mean every backend API call is subsidized. Confirm your eligible billing workspace before you subtract subscription credits from your developer cost projections. The interactive calculator below models gross API consumption and doesn't assume any unverified subscription offsets.
A practical subscription budget and spending controls
Let's walk through a simple example. Say your team subscribes to monthly Premium at its standard $30 list price and turns on a $15 monthly spending cap for paid Additional Usage. Your planned service budget is exactly $45 before tax. If your workloads only consume $6 of metered usage, your combined service consumption is $36, leaving $9 in prepaid credit balance. An explicit spending cap is a cheap safety net against runaway loops, and it keeps API access running without surprises.
Automatic account top-ups can be toggled on or off at any time. If you buy an annual Premium membership at $300 upfront and add a $15 initial usage buffer, your starting out-of-pocket cash is $315. It helps to track cash outflow separately from monthly amortized cost, which keeps your budget accounting clean when finance comes asking.
Grok 4.7 API: calculating gross token consumption
If you're calling xAI models directly through the API, you'll be on the Grok 4.7 endpoints. According to the Grok 4.7 pricing documentation, standard inference costs $2.00 per million uncached input tokens, $0.50 per million cached input tokens, and $6.00 per million output tokens for requests at or below 200K context length. Here's how a few representative workloads work out in gross API spend:
| Workload Description | Input Charge Breakdown | Output Charge Breakdown | Total gross API Cost |
|---|---|---|---|
| 10M uncached input + 2M output | 10 × $2 = $20 | 2 × $6 = $12 | $32.00 |
| 100M uncached input + 20M output | 100 × $2 = $200 | 20 × $6 = $120 | $320.00 |
| 3M uncached + 7M cache-hit + 2M output | $6 + $3.50 = $9.50 | 2 × $6 = $12 | $21.50 |
Look at what context caching does for you here. Swapping seven million uncached input tokens for cached ones saves $10.50 on a single run, which is the kind of win that makes caching worth setting up early. Keep in mind these are gross API estimates, before local sales tax, separate embedding endpoints, and storage fees.
Grok versus ChatGPT and Claude: matching tools to jobs
If you're choosing between Grok, ChatGPT Plus, and Claude Pro for daily knowledge work, the best test is your own tasks. Look at real-time information access, citation transparency, and how consistent the reasoning is on the work your company actually does. For software engineering, compare developer endpoints, function calling reliability, and context caching performance across vendors.
If you want to see the subscription options side by side, our guide ChatGPT Plus vs Claude Pro vs Google AI Pro is a good next stop. For token pricing across open and proprietary models, take a look at our Cheapest AI API Benchmark.
Common buying questions
Do I have to switch my existing SuperGrok plan?
Nope. The provider confirms that existing plan holders can stay on their legacy subscriptions. Before you decide to switch, compare your current invoice and included allowances with the new packages.
Does a Premium plan guarantee coverage for my backend API workload?
No. This guide doesn't assume any automatic token allocation for standalone developer API keys. Check your workspace billing dashboard to be sure. The pricing examples here calculate gross API consumption from published model rates.
Which values should I input into the cost calculator?
Enter your expected uncached input, cached input, and output tokens separately. The interactive worksheet defaults to 10M uncached input tokens and 2M output tokens on Grok 4.7, which gives a $32 gross API total. If your account has promotional credits, subtract them at the final invoice stage rather than changing the underlying model rates.
Calculate gross Grok 4.7 API consumption
Defaults to 10M uncached input plus 2M output tokens at $2/$6 per million: $32 gross usage. Assumes requests at or below 200K context, excludes tools, tax and funding, and applies no unverified subscription credits.
Compare the model rates for your workload →
Sources
- The X Subscription: current USD monthly and annual list prices
- Account unification, existing subscriptions and included products
- Shared weekly usage, eligible surfaces and account limits
- Additional usage, prepaid credit and optional automatic top-ups
- Grok 4.7 API pricing and context conditions
- Current regional checkout and promotional offers
- Existing X Premium features and regional pricing