Home/Grok API pricing

Grok API Pricing

xAI sells Grok 4.7 and Grok 4.3 through its API, and SuperGrok plans for the Grok app. One typical prompt costs 0.19¢ to 0.4¢ depending on the model. Verified October 2026.

Is Grok free?

Grok is free to try on the web and in the apps, with higher limits on SuperGrok plans.

Grok API Prices by Model

ModelInput / 1MOutput / 1MPer prompt
Grok 4.3$1.25$2.50$0.0019
Grok 4.7$2.00$6.00$0.0040
Grok 4.20Older$1.25$2.50$0.0019
Grok 4.1 FastRetired$0.20$0.50$0.000350

Batch: The Batch API is 20% off on Grok 4.3 and Grok 4.20 only; Grok 4.7 does not support batch.

Caching: Cached input is billed at a reduced rate (about 25% of input on Grok 4.7); there is no write charge. Prompts of 200K+ tokens are billed at 2x.

Free API tier: None — the API is paid from the first token.

Source: xAI official pricing · verified October 2026.

Grok Plans

Free
Free

Chat, real-time web and X search, reasoning, image and video generation.

SuperGrok Lite
$10/mo

Longer conversations, Expert mode, the latest model.

SuperGrok
$30/mo
₹2,900/mo

5x longer conversations, the best model with higher limits, 720p video.

SuperGrok Plus
$100/mo

Much higher usage, 1080p video, priority at peak times.

SuperGrok Heavy
$300/mo

Highest usage and speed, Grok Heavy multi-agent, X Premium+.

Subscription or API: which is cheaper?

PlanPrice / monthAPI prompts / day
Grok SuperGrok Lite$10~83Grok 4.7
Grok SuperGrok$30₹2,900~250Grok 4.7
Grok SuperGrok Plus$100~833Grok 4.7
Grok SuperGrok Heavy$300~2,500Grok 4.7

Typical prompt = 500 input + 500 output tokens, priced on the API model the plan says it includes (or the provider’s top model when it does not say). If you would send more prompts than this every day, the subscription is the better deal — within its usage limits. Fewer, and paying per API call is cheaper. Subscriptions also include apps, tools and features the raw API does not.

Plan details: Grok official page.

Compare xAI With Every Other Model

0input tokens
Estimated (~4 chars/token)

How many tokens you expect the model to generate per response

Showing a typical prompt (500 input + 500 output tokens) across 27 models. Paste your own prompt above for an exact count.

ModelPer PromptPrompts / $1Monthly
xAI$0.00190.19¢533$5.63
xAI$0.00190.19¢533$5.63
xAI$0.00400.4¢250$12.00
OpenAI$0.0003000.03¢3,333$0.9000
Mistral$0.0003750.037¢2,666$1.13
OpenAI (via Together)$0.0003750.037¢2,666$1.13
OpenAI$0.0007000.07¢1,428$2.10
DeepSeek$0.0007500.075¢1,333$2.25
Meta (via Together)$0.0009250.092¢1,081$2.77
Mistral$0.00100.1¢1,000$3.00
Google$0.00140.14¢714$4.20
Google$0.00220.22¢444$6.75
DeepSeek$0.00260.26¢378$7.92
Meta$0.00270.27¢363$8.25
Anthropic$0.00300.3¢333$9.00
Alibaba (Qwen)$0.00400.4¢250$12.00
Mistral$0.00450.45¢222$13.50
OpenAI$0.00600.6¢166$18.00
OpenAI$0.00600.6¢166$18.00
Anthropic$0.00600.6¢166$18.00
OpenAI$0.00700.7¢142$21.00
Google$0.00700.7¢142$21.00
Moonshot AI$0.00900.9¢111$27.00
OpenAI$0.012083$36.00
Anthropic$0.012083$36.00
OpenAI$0.030033$90.00
Anthropic$0.030033$90.00

Prices from official provider pricing pages, verified October 2026.

Grok API & Grok Pricing — FAQ

The Grok API is pay-per-token. One typical prompt (500 input + 500 output tokens) costs between $0.0019 and $0.0040 depending on the model, across 3 models available today. Prices were verified against xAI's official pricing page in October 2026.

Grok is free to try on the web and in the apps, with higher limits on SuperGrok plans.

Grok plans: SuperGrok Lite $10/month; SuperGrok $30/month (₹2,900 in India); SuperGrok Plus $100/month; SuperGrok Heavy $300/month.

Grok 4.3 is the cheapest xAI model available: $1.25 per million input tokens and $2.50 per million output tokens, about 0.19¢ per typical prompt — $1 buys about 533 prompts.

The Batch API is 20% off on Grok 4.3 and Grok 4.20 only; Grok 4.7 does not support batch.

Grok SuperGrok costs $30 a month — the price of about 7,500 typical prompts on Grok 4.7 through the API, or roughly 250 a day. If you send more than that, the subscription is cheaper (within its usage limits); if you send fewer, paying per API call costs less.