Plans and pricing

Updated November 2, 2026: Imagine. grok-imagine-image-quality was retired on November 2, 2026: its requests are now served by Imagine Image 2.0 (grok-imagine-image-2.0, $0.04 per image) at low quality. If you have older code, replace the model ID and set the quality you want; grok-imagine-image (version 1.0) isn’t affected. Imagine Image 2.0 accepts up to five reference images and auto quality, now the default.

How xAI pricing works

xAI offers several levels of access to Grok, from the free plan to enterprise offers. They differ less in price than in billing logic: the consumer subscription is paid monthly, while the API is paid for by usage. Understanding that structure avoids the classic mistake of paying for a subscription when you need an integration, or opening an API account for a use that would fit in the free plan.

Plans for individuals

The free plan on grok.com gives access to the basic features: conversations with Grok, web search, image generation within a set limit, and use on the web as well as in the mobile apps. Its restrictions are mainly about volume: a number of messages per day, limited access to the most powerful models, and queues during peak demand. For exploring or occasional use, it’s more than enough; the queue, however, becomes a nuisance as soon as Grok is part of your daily work routine.

The Premium plan lifts those constraints. It offers unlimited use of the most powerful models, removes the queue, unlocks advanced features such as high-quality image generation and video, and gives early access to new features. You manage the subscription on grok.com, or through the Apple and Google app stores for the mobile apps. Subscribing through a store makes payment simpler but puts cancellation under the platform’s rules, which is worth knowing before you choose.

API pricing

On the API, you no longer pay a flat fee but for what you actually use: by the token for text, by the unit for images, and by the second for video.

The gap between those two rows deserves a closer look: grok-4.3 costs about 1.6 times less for input and 2.4 times less for output, while offering the largest context window in the lineup. Routing simple tasks to grok-4.3 and keeping grok-4.7 for problems that genuinely require reasoning is the most effective way to save money, well before any prompt optimization.

Automatic caching also lowers the cost of repetitive requests: on grok-4.3 and grok-4.20, cached input tokens cost $0.20 per million instead of $1.25, about 84% less. It turns on without any configuration, but how well it works depends on you: by passing a conversation ID in your requests, you let the system recognize message sequences it has already seen. On an assistant where every call resends a long system prompt, the difference at the end of the month is dramatic.

Built-in tools are billed separately from tokens, based on the number of calls.

One last lever: the Batch API gives a 50% discount on standard token prices for batch jobs that don’t need an immediate response. Image and video costs stay at the normal rate. An overnight classification of thousands of documents, a catalog translation, a database enrichment: anything that can wait a few hours should go through it.

Billing and payment

You can buy credits in advance from the xAI console; they’re drawn down as you use the API. Two points call for particular care: prepaid credits are non-refundable, and invoices can’t be regenerated, so check your billing information (company name, VAT number, address) before your first purchase, not after. Beyond your credits, additional usage is billed at the end of the month, with real-time tracking available in the console. Check it from the start of the project, when volumes are low and design mistakes are cheap to fix.

Grok for Business

For companies, xAI offers custom deals through the Grok for Business program; requests go through the form at x.ai/grok/business or by email. These offers generally include volume-based negotiated pricing, dedicated support, service-level agreements, and advanced compliance options such as HIPAA or SOC 2. That last point is often decisive: in healthcare or finance, compliance requirements, more than the price per token, are what push you toward this channel.

Prices as of August 5, 2026. Prices change often: before you budget anything, check the current rates on the official xAI models and pricing page.

Key takeaways

• The free plan on grok.com offers basic access to Grok
• The API offers transparent pricing by the token, the image, or the second
• Automatic caching cuts the cost of repeated input tokens by 75 to 85%, depending on the model
• The Batch API gives a 50% discount for non-urgent jobs
• Prepaid credits are non-refundable
• Grok for Business offers custom deals for companies