Grok API pricing is $3 per 1M input tokens and $15 per 1M output on Grok 4,
which blends to roughly $5.40 at a typical usage mix. That places xAI in the same bracket as
frontier models from OpenAI and Anthropic rather than in the budget tier — a positioning
decision that shapes when Grok is the right call and when it is not.
Grok API pricing: every model, per 1M tokens
| Model | Input $/1M | Output $/1M | Blended $/1M | Context |
|---|---|---|---|---|
| Grok 4 | $3.00 | $15.00 | $5.40 | 256,000 tokens |
Blended is the effective rate at a 4:1 input-to-output mix, which is what a typical chat or retrieval workload actually produces. It is the number to compare across vendors — a headline input price hides how much the output side costs.
What Grok costs per month
| Workload | Tokens / month | Grok 4 |
|---|---|---|
| Side project | 1M in / 0.25M out | $6.75 |
| Small team | 20M in / 5M out | $135 |
| Production | 200M in / 50M out | $1,350 |
Model your own volumes in the AI API cost calculator.
How Grok compares to other providers
Each provider's cheapest priced model, so the comparison is like for like on entry cost.
| Provider | Cheapest model | Blended $/1M |
|---|---|---|
| Mistral AI | Mistral 7B | $0.0220 |
| Meta | Llama 3.1 8B | $0.0220 |
| Alibaba | Qwen3 8B | $0.0600 |
| Gemma 3 4B | $0.0600 | |
| Microsoft | Phi-4 | $0.0840 |
| DeepSeek | DeepSeek V4-Flash | $0.168 |
| Moonshot AI | Kimi K2.7 Code | $0.980 |
| Anthropic | Claude Haiku 4.5 | $1.80 |
| Zhipu AI | GLM 5.2 | $2.00 |
| xAI this page | Grok 4 | $5.40 |
| OpenAI | GPT-5.6 Sol | $10.00 |
Cheapest is not the same as best value — check capability alongside price on the LLM leaderboard, or browse every model in the AI models database.
What you get for the price
Grok 4 launched on 9 July 2025 as xAI’s flagship reasoning model and the successor to
Grok 3. It pairs native chain-of-thought reasoning with multimodal understanding, accepting
text and image input and returning text, and ships with a 256,000-token context window. At
release it posted strong results on reasoning and tool-use benchmarks including Humanity’s Last
Exam and ARC-AGI, helped by real-time access to X and the web.
That live-search access is the genuine differentiator. Most frontier models answer from
training data plus whatever you put in the prompt; Grok can reach current information directly.
For workloads where recency is the requirement — monitoring, research on breaking topics,
anything where a stale answer is a wrong answer — that is worth paying for in a way that
benchmark scores do not capture.
Access is entirely proprietary. There are no open weights, and the model runs through the
xAI API, the Grok apps and the SuperGrok subscription. xAI has since shipped faster, lower-cost
variants of Grok 4; we list published per-token pricing only for models we have verified, so
this page covers Grok 4 itself.
Where the 256K context bites
The context window is the clearest limitation against direct rivals. Grok 4 offers 256,000
tokens where Claude, Gemini and the DeepSeek V4 line all reach roughly 1M. For most
applications 256K is more than sufficient — it is a large book. But for whole-repository
reasoning, long document archives or workloads that deliberately avoid a retrieval layer, it
is a real constraint at a price point where competitors do not impose one.
The counterweight is that Grok has no long-context premium. GPT-5.6 Sol bills 2x input and
1.5x output above 272K tokens; Gemini 3.1 Pro roughly doubles input above 200K. Grok’s rate is
flat across its whole window, so what you see is what you pay.
Is Grok worth $5.40 blended?
It depends almost entirely on whether you need live search. Judged purely on capability per
dollar, the comparison is unflattering: the cross-provider table above shows several models at a
fraction of the cost, and open-weight flagships now score within a few points of the best closed
models while charging far less. A team choosing on price alone would not land on Grok.
Judged on what it uniquely does, the calculation changes. Real-time access to X and the
broader web is not a feature you can replicate by paying a cheaper provider more tokens — you
would have to build and maintain a retrieval pipeline, which has its own cost and its own
failure modes. If that capability is central to your product, Grok’s pricing is reasonable for
the bracket it sits in.
If it is not central, treat Grok as a specialist you route specific requests to rather than
a default. Estimate what that split would cost in the
AI API cost calculator, and compare
capability against price for every model in the
LLM leaderboard before committing.
Frequently asked questions
How much does the Grok API cost?
Grok costs $3.00 per 1M input tokens and $15.00 per 1M output on Grok 4, which blends to $5.40 per 1M at a typical 4:1 input-to-output mix. Input and output are billed separately, and output is the more expensive side.
How much does Grok cost per month?
At 20M input and 5M output tokens a month, Grok 4 costs about $135. A side project at 1M input and 0.25M output costs a small fraction of that. Use the AI API cost calculator for your own volumes.
Is Grok cheaper than Mistral AI?
On entry-level pricing, Grok starts at $5.40 per 1M blended and Mistral AI starts at $0.0220 on Mistral 7B. Mistral AI is the cheaper entry point, though capability differs — compare both on the leaderboard before switching.
Why does Grok charge more for output tokens than input?
Output tokens are generated one at a time and cannot be batched the way a prompt can, so they cost more to serve. This is why prompt-heavy workloads such as retrieval and classification are far cheaper to run than generation-heavy ones, and why the blended rate matters more than the headline input price.
Prices are the published list rates for each model's primary API and are reviewed as providers change them. Volume, batch and cached-input discounts are not included. Last reviewed August 2026.
