Grok 4.7 API
NewLLMThe Grok 4.7 API serves xAI's newest flagship coding model here, with a 500K context window, vision, and structured tool use.
- Input · USD / 1M tokens$3.2 USDOfficial price
$4 USD - Output · USD / 1M tokens$9.6 USDOfficial price
$12 USD - Savings20%
- Context500K
- View tool pricing and billing rules
What are we making today?
Bring an idea, a question, or a first draft.
01
What this model is for
The Grok 4.7 API serves xAI's newest flagship coding model here, with a 500K context window, vision, and structured tool use.
Grok 4.7 is xAI's current flagship in the three-strong Grok family listed on OmniaKey, released on September 21, 2026 and marked as new and featured in the catalog. It accepts text and images, exposes a 500K context window, supports tool use, and carries the same input, output and cache-read listing as Grok 4.6, so the upgrade sits in behaviour rather than in a bigger window or a cheaper token.
xAI describes 4.7 as a new, larger base model trained with a longer reinforcement-learning run and a mix weighted toward tasks that take hours. That makes it the first id to try for a coding agent working through an unfamiliar repository, and 4.6 the id to keep for a pinned deployment or an A/B check. Grok 4.5 stays listed as the older revision with a lower cache-read rate, which suits a loop that rereads one large, stable prompt.
Best-fit coding workloads
Long repository refactors
Give it a checkout, a failing test and the intended change, then let it work through several files before you read the diff. The 500K window carries the tool output of a long session, and reasoning effort can be pinned to high on both generations when you compare revisions.
Reasoning state across turns
On the Responses API this id returns reasoning.encrypted_content even when include does not request it. Send those items back unchanged on the next turn, so a multi-turn agent keeps its reasoning state instead of rebuilding it from the visible transcript alone.
Screenshot-led debugging
Hand it a UI screenshot, the error text and the behaviour you expected, then ask which element is wrong and why. Vision and tool use are both listed, so an image and a log excerpt can sit in one thread. Ask it to point at the region it read, and check that region yourself.
02
Hands-on notes
Checked on 2026-10-27 through the OmniaKey API with one standard coding prompt (median of 3 runs) and three capability probes.
- Output tokens for our standard coding prompt, reasoning included
- 2,597
- Cost of that output at current prices
- $0.0249
- JSON mode
- Passed
- Tool calling
- Passed
- Image input
- Passed
03
Technical specification
Catalog facts and integration details verified for this route.
- Model ID
- grok-4.7
- Context
- 500K
- Inputs and tools
- Vision · Tool use
- API endpoint
- /v1/chat/completions
- Cache hit / 1M
- $0.8 USD
- Coding clients
- Codex · Cursor · Cline · Aider
Limits to plan around
- The 500K figure is prompt capacity, not an output allowance. The catalog lists no maximum output length for this id, so a long generation has to be planned around what the reply actually returns.
- Batch is not offered for this id, so a nightly evaluation job cannot be moved into a batch queue that prices the same tokens differently.
- xAI's launch comparison runs 4.7 at xhigh and 4.6 at high, so swapping ids without pinning the same effort changes the reasoning budget as well as the model.
- The Fast variant is sold only inside Cursor and Grok Build. A request to the public API or to OmniaKey uses the standard id, so a Fast name will not resolve.
04
Compare nearby models
Compare workload fit and current listed token rates before choosing a route.
| Model | Best for | Context | Input | Output |
|---|---|---|---|---|
| Grok 4.7This page grok-4.7 | Newest Grok revision; first pick for long agent threads that want xAI's current model. | 500K | $3.2 USD Official price $4 USD | $9.6 USD Official price $12 USD |
| Grok 4.6 grok-4.6 | Pinned 4.6 deployments and revision A/B checks; same window and same listed rates as 4.7. | 500K | $3.2 USD Official price $4 USD | $9.6 USD Official price $12 USD |
| Grok 4.5 grok-4.5 | Older revision with the lower cache-read rate, for loops that reread one large stable prompt. | 500K | $3.2 USD Official price $4 USD | $9.6 USD Official price $12 USD |
05
SDK quickstarts
Use one base URL and model ID from cURL, Python, or TypeScript.
curl https://api.omniakey.com/v1/chat/completions \
-H "Authorization: Bearer $OMNIAKEY_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "grok-4.7",
"messages": [
{"role": "user", "content": "Review this pull request"}
],
"stream": true
}'06
Estimate token cost
Enter expected input and output tokens to compare current listed rates.
- OmniaKey estimate
- $0.512
- Official-rate estimate
- $0.64
- Estimated savings
- $0.128
Estimate covers input and output tokens only. Cache usage, retries, tools, and client-side token accounting can change the final charge.
07
Related coding guides
Deeper selection, setup, and troubleshooting guidance for this model family.
Grok 4.7 Review
Grok 4.7 is a credible same-price upgrade over 4.6 for long coding and knowledge-work agents, but xAI's launch scores are not a substitute for a matched workload test.
Read guideGrok 4.6 API Pricing
xAI lists Grok 4.6 at $2 input and $6 output per million tokens below the long-context threshold, then doubles those rates for prompts above 200K tokens.
Read guide08
Sources and verification
Model facts and links were checked on 2026-10-27. Provider documentation remains the authority for upstream behavior.
Frequently asked questions
How much does Grok 4.7 cost on OmniaKey?
OmniaKey charges $3.2 USD per 1M input tokens and $9.6 USD per 1M output tokens, in USD. Official prices are $4 USD and $12 USD; you save 20%.
What is the context window for Grok 4.7?
The listed context window is 500K. Actual usable limits can also depend on the client and upstream request format.
How do I call Grok 4.7 through OmniaKey?
Send a POST to https://api.omniakey.com/v1/chat/completions with grok-4.7 as the model field and an OmniaKey key as the bearer token. The catalog lists codex, cursor, cline and aider as supported clients, each configured with the same base URL and key. Model ids are case-sensitive, so copy the spelling exactly.
Is Grok 4.7 better than Grok 4.6?
xAI reports higher results for 4.7 than 4.6 across the task families in its launch comparison, and both ids carry the same window and the same listed rates. That comparison runs the two at different effort settings, so reproduce a few of your own tasks with effort pinned before you move a pinned 4.6 deployment over.
How large is the Grok 4.7 context window?
The catalog lists 500K tokens as the context window, and it covers the prompt: images, tool output and the running transcript all draw on the same allowance. The catalog lists no maximum output length for this id, so a long tool loop should trim stale tool output before it extends a session, and the reply itself is the only ceiling in view.
Does Grok 4.7 accept images and tool calls?
Yes. The catalog lists vision and tool use, and xAI's start guide names function calling, structured output, web search, X search and code execution for this id. Images and tool results can share one thread, which is what makes screenshot-led debugging workable inside a coding agent.
What is different about Grok 4.5?
Grok 4.5 is the older revision: the same 500K window, the same vision and tool support, and a lower cache-read rate than 4.6 and 4.7. Pick it when a thread rereads a large stable prompt often enough for that line item to matter, and test behaviour first, because neighbouring revisions of one family rarely fail in the same way.
How is Grok 4.7 billed on OmniaKey?
Usage is charged to your account balance per input and output token, with cached input billed at its own separate rate. The current rates are in the price table on the model page, and one key covers this model and every other id in the catalog, so a single balance pays for all of them.