Real Costs: Free, API, and Tiered Pricing
On this page
Updated November 2, 2026 — flagship model. Grok 4.5 is no longer the most capable model: Grok 4.6 (August 12, 2026) and then Grok 4.7 (
grok-4.7, September 21, 2026) replaced it, with 500,000 context tokens, reasoning effort fromlowtoxhigh, and the same tiered pricing ($2 / $6 under 200,000 prompt tokens, $4 / $12 above).grok-4.5remains available; for a new project, start withgrok-4.7(course Grok in 2026, lesson 9).
Understand Grok Pricing
Before you build a tool into the core of your operations, you need to know exactly what it will cost. Otherwise, the budget line drifts without anyone noticing, and you discover the following quarter an expense you can neither explain nor allocate. This lesson details Grok’s pricing structure so you can anticipate your spending, decide which model to use for which job, and bill clients properly for what needs billing.
Free access via grok.com
The grok.com conversational interface is accessible for free. You ask questions and get complete answers, use integrated web and X search, attach documents for analysis, and generate images, all without spending a cent. The free plan imposes limits on daily requests, but it remains largely sufficient for moderate individual use, on the order of a few dozen requests per day. A consultant who writes reports, prepares two proposals, and checks industry news won’t hit this ceiling on an ordinary day.
API pricing: pay-as-you-go
The API runs on a consumption model: you pay only for what you use, with no subscription or commitment. Text model rates calculate per million tokens, which represents about 750,000 words. This order of magnitude helps you get your bearings: a professional email weighs barely a few hundred tokens, so you need considerable volume before the bill becomes noticeable.
| Model | Input ($/M tokens) | Output ($/M tokens) | Recommended usage |
|---|---|---|---|
| grok-4.5 (flagship) | $2.00 | $6.00 | Complex tasks, deep analysis |
| grok-4.3 | $1.25 | $2.50 | Routine tasks, high volume |
Integrated tool calls add to the tokens, billed separately. Each triggered search counts, which means you should avoid activating web search on queries that don’t need it: rewriting an email doesn’t require querying the web, and automation that does this systematically pays for nothing.
| Billed element | Rate |
|---|---|
| Web Search | $5 per 1,000 calls ($0.005 per search) |
| X Search | $5 per 1,000 calls ($0.005 per search) |
| Code execution | $5 per 1,000 calls |
| Attachments | $10 per 1,000 calls |
| Standard image | $0.02 per image |
| High-quality image | $0.05 per image |
| Video | $0.05 per second |
What this looks like for a typical SMB
These unit prices mean little until you relate them to real usage. Let’s take a freelance consultant who uses Grok every day to write, analyze, and monitor news.
| Usage | Monthly volume | Estimated cost |
|---|---|---|
| Email and report writing | 200 requests (fast model) | ~$2 |
| Document analysis | 50 requests (standard model) | ~$3 |
| X Search monitoring | 100 searches | ~$0.50 |
| Web Search monitoring | 100 searches | ~$0.50 |
| Monthly total | ~$6 |
For a five-person SMB with significantly higher volume, the monthly cost usually sits between $20 and $80 depending on usage intensity. Compared to the time saved, the order of magnitude is incomparable to traditional business software, billed by subscription and per seat.
Two mechanisms that lower the bill
Grok automatically applies a cache to requests. Tokens already processed in a conversation are billed at the “cached” rate: on grok-4.5, a token billed $2.00 per million on input drops to $0.30 (85% less), and on grok-4.7 to $0.50 (75% less). You have nothing to configure; the mechanism triggers as soon as you maintain a stable conversation context. In practice, the fifteen-page product documentation that your chatbot sends with every request is paid at full rate only once.
The Batch API is the second lever. If your tasks aren’t urgent—nightly report generation, batch file analysis—it offers a 20% reduction on tokens for eligible models (from grok-4.3 to grok-4.7 as of November 2, 2026, according to xAI batch pricing). Processing is asynchronous: you submit your requests and retrieve results later, generally within 24 hours. A firm that prepares nightly summaries of the previous day’s requests has no reason to pay real-time rates for a document no one will read for eight hours.
Billing and payment
The system runs on prepaid credits: you buy an amount in advance and consume from it. Beyond these credits, monthly billing takes over at month’s end. Your invoices are available in the xAI console, at console.x.ai, which facilitates transmission to your accountant. Remember finally that prepaid credits are non-refundable: start with a modest amount, measure your real consumption over a few weeks, then recharge knowing what you need.
Prices as of August 5, 2026 — prices evolve regularly: before any budget calculation, check the current rate card on the official xAI models and pricing page.
Key takeaways
- grok.com is free for standard daily use
- The API bills per token consumed, no subscription
- The economy model (grok-4.3) costs about half as much as the flagship model
- Automatic cache and Batch API allow significant bill reduction
- A typical freelancer spends between $5 and $15 per month via the API