Grok API pricing (xAI)
As of November 3, 2026, the Grok API costs $1 to $2 per million input tokens and $2 to $6 per million output tokens, across 7 Grok models. Grok Build 0.1 costs $1 input and $2 output; Grok Code Fast costs $1 input and $2 output; Grok 4.7 costs $2 input and $6 output. A typical request of 1,000 input and 500 output tokens costs $0.002 on Grok Build 0.1 and $0.005 on Grok 4.7.
Grok API price per 1M tokens, every model
| Model | Input | Cached input | Output | Batch input | Batch output | Context | Typical request |
|---|---|---|---|---|---|---|---|
Grok Build 0.1grok-build-0.1 | $1 | $0.2 | $2 | – | – | 256K | $0.002 |
Grok Code Fastgrok-code-fast | $1 | $0.2 | $2 | – | – | 256K | $0.002 |
Grok 4.3grok-4.3 | $1.25 | $0.2 | $2.5 | $1 | $2 | 1M | $0.0025 |
Grok 4.20grok-4.20 | $1.25 | $0.2 | $2.5 | $1 | $2 | 1M | $0.0025 |
Grok 4.5grok-4.5 | $2 | $0.3 | $6 | – | – | 500K | $0.005 |
Grok 4.6grok-4.6 | $2 | $0.5 | $6 | – | – | 500K | $0.005 |
Grok 4.7grok-4.7 | $2 | $0.5 | $6 | $1.6 | $4.8 | 500K | $0.005 |
US dollars per million tokens, sorted by the cost of a typical request of 1,000 input and 500 output tokens. Also listed as grok-code-fast-1-0825 (same as Grok Code Fast), grok-4.20-0309-non-reasoning (same as Grok 4.20), grok-4.20-0309-reasoning (same as Grok 4.20), grok-4.20-non-reasoning (same as Grok 4.20) and grok-4.20-reasoning (same as Grok 4.20).
How much does a Grok API request cost?
A typical request of 1,000 input and 500 output tokens costs $0.002 on Grok Build 0.1, $0.002 on Grok Code Fast and $0.005 on Grok 4.7. A coding agent step with 50K tokens of context and 2K output tokens costs $0.11 on Grok 4.7.
| Model | Short chat message 300 in, 200 out | Typical request 1K in, 500 out | Long document summary 20K in, 1K out | Coding agent step 50K in, 2K out | 1,000 typical requests |
|---|---|---|---|---|---|
| Grok Build 0.1 | $0.0007 | $0.002 | $0.022 | $0.054 | $2.00 |
| Grok Code Fast | $0.0007 | $0.002 | $0.022 | $0.054 | $2.00 |
| Grok 4.7 | $0.0018 | $0.005 | $0.046 | $0.11 | $5.00 |
At standard rates, for the newest model of each Grok family. A token is about 4 characters of English text, so 1,000 tokens is roughly 750 words. xAI counts tokens with its own tokenizer, so the same text can be a different number of tokens.
How batch, prompt caching and long context change the price
Batch API. 3 of the 7 priced Grok models have a batch API price. Requests sent as a batch are processed within 24 hours and cost 20% less: Grok 4.7 drops from $2 input and $6 output to $1.6 and $4.8 per million tokens. See batch API pricing.
Prompt caching. All 7 priced Grok models charge less for cached input, the part of a prompt that repeats across requests: typically 80% less. Grok 4.7 charges $0.5 per million cached input tokens instead of $2. See prompt caching pricing.
Long context. Grok models accept 256K to 1M input tokens. Long prompts cost more on 7 models: Grok Build 0.1 charges $2 input and $4 output per million tokens once a prompt is over 200K tokens, instead of $1 input and $2 output; Grok Code Fast charges $2 input and $4 output per million tokens once a prompt is over 200K tokens, instead of $1 input and $2 output; Grok 4.3 charges $2.5 input and $5 output per million tokens once a prompt is over 200K tokens, instead of $1.25 input and $2.5 output; Grok 4.20 charges $2.5 input and $5 output per million tokens once a prompt is over 200K tokens, instead of $1.25 input and $2.5 output; Grok 4.5 charges $4 input and $12 output per million tokens once a prompt is over 200K tokens, instead of $2 input and $6 output; Grok 4.6 charges $4 input and $12 output per million tokens once a prompt is over 200K tokens, instead of $2 input and $6 output; Grok 4.7 charges $4 input and $12 output per million tokens once a prompt is over 200K tokens, instead of $2 input and $6 output. A 300K-token prompt with 2K output tokens on Grok 4.7 costs $1.22, against $0.61 at the standard rate.
Grok API vs OpenAI and Claude
It depends on the model. The cheapest Grok model, Grok Build 0.1, costs $0.002 per typical request; the cheapest OpenAI model, GPT-5 nano, costs $0.00025. Grok 4.7 ($0.005) is closest in price to GPT-4.1 ($0.006): 1.2× the price.
| Grok model | Input / output | Typical request | Closest OpenAI model | Closest Claude model |
|---|---|---|---|---|
| Grok Build 0.1 | $1 / $2 | $0.002 | GPT-5.4 mini $0.75 / $4.5 · $0.003 · 1.5× the price | Claude Haiku 4.5 $1 / $5 · $0.0035 · 1.8× the price |
| Grok Code Fast | $1 / $2 | $0.002 | GPT-5.4 mini $0.75 / $4.5 · $0.003 · 1.5× the price | Claude Haiku 4.5 $1 / $5 · $0.0035 · 1.8× the price |
| Grok 4.7 | $2 / $6 | $0.005 | GPT-4.1 $2 / $8 · $0.006 · 1.2× the price | Claude Sonnet 5 $2 / $10 · $0.007 · 1.4× the price |
For each Grok model, the model from the other provider whose typical request costs the nearest, with its input / output price per million tokens, its typical request, and how that compares with the Grok model.
Questions and answers
How much does the Grok API cost?
As of November 3, 2026, the Grok API costs $1 to $2 per million input tokens and $2 to $6 per million output tokens, across 7 Grok models. Grok Build 0.1 costs $1 input and $2 output; Grok Code Fast costs $1 input and $2 output; Grok 4.7 costs $2 input and $6 output. A typical request of 1,000 input and 500 output tokens costs $0.002 on Grok Build 0.1 and $0.005 on Grok 4.7.
What is the cheapest Grok model?
Grok Build 0.1 (grok-build-0.1), at $1 per million input tokens and $2 per million output tokens.
How much does a Grok API request cost?
A typical request of 1,000 input and 500 output tokens costs $0.002 on Grok Build 0.1, $0.002 on Grok Code Fast and $0.005 on Grok 4.7. A coding agent step with 50K tokens of context and 2K output tokens costs $0.11 on Grok 4.7.
Is the Grok batch API cheaper?
Yes. 3 of the 7 priced Grok models have a batch API price. Requests sent as a batch are processed within 24 hours and cost 20% less: Grok 4.7 drops from $2 input and $6 output to $1.6 and $4.8 per million tokens.
How much does prompt caching save on the Grok API?
All 7 priced Grok models charge less for cached input, the part of a prompt that repeats across requests: typically 80% less. Grok 4.7 charges $0.5 per million cached input tokens instead of $2.
Does long context cost more on the Grok API?
Yes, on some models. Grok models accept 256K to 1M input tokens. Long prompts cost more on 7 models: Grok Build 0.1 charges $2 input and $4 output per million tokens once a prompt is over 200K tokens, instead of $1 input and $2 output; Grok Code Fast charges $2 input and $4 output per million tokens once a prompt is over 200K tokens, instead of $1 input and $2 output; Grok 4.3 charges $2.5 input and $5 output per million tokens once a prompt is over 200K tokens, instead of $1.25 input and $2.5 output; Grok 4.20 charges $2.5 input and $5 output per million tokens once a prompt is over 200K tokens, instead of $1.25 input and $2.5 output; Grok 4.5 charges $4 input and $12 output per million tokens once a prompt is over 200K tokens, instead of $2 input and $6 output; Grok 4.6 charges $4 input and $12 output per million tokens once a prompt is over 200K tokens, instead of $2 input and $6 output; Grok 4.7 charges $4 input and $12 output per million tokens once a prompt is over 200K tokens, instead of $2 input and $6 output. A 300K-token prompt with 2K output tokens on Grok 4.7 costs $1.22, against $0.61 at the standard rate.
Is the Grok API cheaper than OpenAI?
It depends on the model. The cheapest Grok model, Grok Build 0.1, costs $0.002 per typical request; the cheapest OpenAI model, GPT-5 nano, costs $0.00025. Grok 4.7 ($0.005) is closest in price to GPT-4.1 ($0.006): 1.2× the price.
Source: xAI pricing, via the LiteLLM price list.