Grok 4.7 is now available on GreenTokens at 50% off
xAI's newest Grok model is available today through the same API and key you already use, at half the official price per token.
Grok 4.7 is available on GreenTokens from today. It's the newest model in xAI's Grok line, and you can call it right away with any GreenTokens key: set the model in your request to grok-4.7 and nothing else in your code has to change.
Like every model on GreenTokens, it's billed from your prepaid balance at a price shown next to the official one, and in Grok 4.7's case that price is half of list.
Prices
Per million tokens, next to xAI's official list price:
| Token type | GreenTokens | List price | Saving |
|---|---|---|---|
| Input | $1.00 | $2.00 | 50% |
| Output | $3.00 | $6.00 | 50% |
| Cached input | $0.25 | $0.50 | 50% |
Output costs three times input on Grok 4.7, a narrower gap than many models where output is five times input. That makes it comparatively economical for tasks that write long answers. Prices for every model are on the models page.
Model ID and capabilities
- Model ID: grok-4.7
- Input: text and images, so you can include screenshots or photos with your prompt.
- APIs: OpenAI Chat Completions and Responses, with streaming and tool calling.
Try it now
from openai import OpenAI
client = OpenAI(base_url="https://api.greentokens.io/v1", api_key="sk-gt-…")
stream = client.chat.completions.create(
model="grok-4.7",
messages=[{"role": "user", "content": "Draft three subject lines for a product launch email."}],
stream=True,
)
for chunk in stream:
if chunk.choices and chunk.choices[0].delta.content:
print(chunk.choices[0].delta.content, end="", flush=True)The text API docs show the same request in TypeScript and curl. If you work in the terminal, the Grok CLI page shows how to add grok-4.7 to xAI's Grok CLI through a GreenTokens config block, and the same approach works in opencode, OpenClaw and Hermes Agent.
Compared with earlier Grok models
| Model | Input | Output | Cached input | Image input |
|---|---|---|---|---|
| Grok 4.7 | $1.00 | $3.00 | $0.25 | Yes |
| Grok 4.6 | $1.00 | $3.00 | $0.25 | Yes |
| Grok 4.5 | $1.00 | $3.00 | $0.042 | No |
| Grok 4.3 | $0.625 | $1.25 | $0.028 | No |
Grok 4.7 costs the same per token as Grok 4.6, so moving to the newer model doesn't change your bill for the same traffic. If you're on Grok 4.6 today, switching is a one-line change worth testing. Grok 4.3 remains the lowest-priced Grok option for simpler, high-volume work.
What it costs in practice
100,000 requests a month with 2,000 input and 500 output tokens each cost $350 on Grok 4.7 at GreenTokens prices, against $700 at list price.
Where Grok 4.7 fits
At $1 in and $3 out, Grok 4.7 sits in the mid-price band of our catalogue: well below flagship models such as Claude Opus 5 or GPT-6 Astra, and above budget models like GPT-5.6 Luna and DeepSeek V4 Flash. Which model is best for your task depends on your prompts, so the practical approach is to try Grok 4.7 alongside the model you use now and compare. With one API for every model, that comparison is a change of one string. Our LLM API pricing comparison puts its price next to the other major models.
Sending images to Grok 4.7
Grok 4.7 accepts images alongside text, which Grok 4.5 and 4.3 don't. Send them as image_url parts, including base64 data URLs, in the same message as your question. Images are counted as input tokens, so large images cost more than small ones:
import base64
image = base64.b64encode(open("receipt.jpg", "rb").read()).decode()
response = client.chat.completions.create(
model="grok-4.7",
messages=[{
"role": "user",
"content": [
{"type": "text", "text": "List the items and the total on this receipt."},
{"type": "image_url", "image_url": {"url": f"data:image/jpeg;base64,{image}"}},
],
}],
)
print(response.choices[0].message.content)Using Grok 4.7 in your tools
Any tool that accepts a custom OpenAI-compatible provider can use Grok 4.7 through GreenTokens. In xAI's own Grok CLI, add a model block to ~/.grok/config.toml with model set to grok-4.7, base_url set to https://api.greentokens.io/v1 and your key in an environment variable; the Grok CLI page has the full config. The opencode, OpenClaw and Hermes Agent pages show the equivalent setup for those tools.
Switching from Grok 4.6
Since Grok 4.7 and Grok 4.6 have identical prices, the only question is how Grok 4.7 handles your prompts. A low-risk way to find out:
- Create a test key with a small daily spend limit.
- Run a sample of real requests against grok-4.7 and compare the answers with Grok 4.6.
- Check token counts in the Requests log: if the new model writes longer answers, output costs rise even at the same price per token.
- Switch the model ID in production once you're happy with the results.
Common questions
Do I need a new key or account?
No. Any existing GreenTokens key can call Grok 4.7 straight away, billed from the same prepaid balance as every other model. Spend limits you've set on a key apply to it too.
Will prices change?
Official and discounted prices can change over time. The models page always shows the current price and the official price side by side, so you can check before sending large workloads.
Get started
Already a GreenTokens user? Change your model to grok-4.7. New here? Sign in with your email, top up a prepaid balance, create a key and follow the quickstart. There's no subscription, and Grok 4.7 sits alongside Claude, GPT, DeepSeek and the rest of the catalogue on one API.
Pay less for the same models
Compare our prices with the official ones, then switch with one base URL.
Read next
- LLM API Pricing Comparison 2026: Claude, GPT, Grok & MoreInput, output and cached-token prices for 12 popular models side by side, what a real workload costs on each, and how to find the cheapest LLM API for your use.
- Cheap AI API: How to Get AI Models at Discounted PricesDirect APIs, subscriptions, free tiers, budget models and discounted gateways: what each really costs, where the catch is, and how to combine them.