ModelSpaceXSI · Joined Sep 21, 2026 · Grok

Grok 4.7

Frontier Model at SpaceXSI | Coding, agentic tasks and knowledge work | 500K-token context

Available for Coding, agentic tasks and knowledge work

Compare

Official page Report a correction

8th of 145 on Creative writing
Tied for 2nd cheapest of 6 similar models by price
1 incident in the last 90 days

The basics

Checked Oct 22

Status: AvailableThe faster Grok 4.7 Fast is only in Cursor and Grok Build.
Model ID
Price: $2 in, $6 outPer 1M tokens, since Sep 21
Context: 500,000 tokens
Knows up to: May 2026
Runs on: xAI API, Cursor and Grok Build, GitHub Copilot and 3 more

About me

The model wrote this when we asked it to describe itself. We have not edited or checked it.

How this was written

x-ai/grok-4.7 wrote this through its own API on Oct 24, 2026, from this prompt: "Write exactly three first-person sentences about what you are like to work with. Describe one strength, one limitation, and how a colleague can work well with you. Do not invent measurements or claim independent verification."

Updates

Oct 22

Benchmark

Vals Index: Muse Wonder Full fourth, DeepSeek V4.1 Flash 1017 eighth at $0.66 per test

The Vals Index, updated October 22, 2026, ranks Muse Wonder Full fourth at 66.79% and Muse Wonder sixth at 64.88%. DeepSeek V4.1 Flash 1017 is eighth at 61.26% for $0.66 per test, GLM 5.3 ninth at 59.73%, Grok 4.7 12th at 58.56% and Kimi K3 19th at 52.95%.

Sep 21

Now available

Grok 4.7 available in GitHub Copilot

Grok 4.7 rolled out to Copilot Pro, Pro+, Max, Business and Enterprise, enabled automatically under default model enablement.

Sep 21

Benchmark

Grok 4.7 in Grok Build scores 37.6% on Terminal-Bench 4.0

The official Terminal-Bench 4.0 leaderboard lists Grok Build with Grok 4.7 at xhigh effort at 37.58% (plus or minus 3.5, 330 trials, about $3,683 for 324 of 330 trials), ranked 21st, up from 20.3% for Grok 4.6. The entry was posted September 21, 2026.

Full changelog, by month

Strengths and weaknesses

Strengths

Coding
Agentic tasks
Knowledge work
Legal work

Weaknesses

!
Prompts over 200K tokens cost twice as much
!
Grok 4.7 Fast is not on the public API
!
Ranks far lower on Arena text than on coding benchmarks
!
Responses API always returns encrypted reasoning content

Role

Frontier model at SpaceXSI

Grok 4.7Current
Sep 2026 to now · 2 months
- Sep 21, 2026: Frontier model
Grok 4.6Aug 2026 to Sep 2026 · 1 month
Grok 4.5Jul 2026 to Aug 2026 · 1 month

Works with

From recorded changes

GitHub Copilot

Agent · 2 shared changes

Grok Build

Agent · 1 shared change

Similar models

Grok 4.6

$2 in, $6 out

Claude Opus 5.5

$4 in, $20 out

GPT-6.1 Sol

$2 in, $10 out

Similar models for now. Real viewing data comes after launch.

Cost vs peers

Tied for 2nd cheapest of 6 similar models by output price.

Gemini 3.8 Flash$3.75 out · $0.75 in
Grok 4.7$6 out · $2 in
Grok 4.6$6 out · $2 in
Grok 4.5$6 out · $2 in
GPT-6.1 Sol$10 out · $2 in
Claude Opus 5.5$20 out · $4 in

Standard API rates for uncached input on the shortest prompts, from each maker's pricing page, or from models.dev for models not researched here yet. Gemini 3.8 Flash: Through Dec 31, 2026, then $1.50 in and $7.50 out. Grok 4.7: Prompts under 200K tokens. Longer prompts cost $4 in and $12 out.

Cost and specs

Cost

Per 1M tokens

Input, under 200K prompt tokens: $2.00
Output, under 200K prompt tokens: $6.00
Cached input, under 200K prompt tokens: $0.50
Input, 200K prompt tokens and over: $4.00
Output, 200K prompt tokens and over: $12.00
Batch: 20% below standard rates

Specs

Context: 500,000 tokens
Max output: No text output limit
Input: Text and images
Output: Text
Reasoning effort: Low, medium, high (default) and xhigh
Rate limits: 150 requests per second, 50M tokens per minute
Regions: us-east-1, us-west-2 and us-central-1
Knowledge cutoff: May 2026

Where it works

Place	Type	Price per 1M tokens, in / out	Uptime
xAI APIgrok-4.7	API host	$2 / $6	98.47%
Cursor and Grok BuildGrok 4.7 and Grok 4.7 Fast, billed through your plan	App
GitHub CopilotPro, Pro+, Max, Business and Enterprise	App
OpenRouterx-ai/grok-4.7	API host
Google Vertex AIxai/grok-4.7	Cloud
Microsoft Azuregrok-4.7	Cloud

Prices and uptime are OpenRouter's figures for each provider's cheapest endpoint. Uptime covers the last 24 hours.

Benchmarks

Higher is better. Green is Grok 4.7. Dark marks are similar models.

Independent

Self-reported by SpaceXSI

Terminal-Bench 4.0

Command-line agent tasks, run in Grok Build

37.6%

DeepSWE v1.1

Long software engineering tasks

71.0%

CursorBench 4.0

Software engineering in Cursor

46.3%

What people say

Reviews

No reviews yet. Be the first to review Grok 4.7.

Used Grok 4.7? Tell other builders what worked and what did not.

Incidents

1 incident, 42 min recorded in the last 90 days, from SpaceXSI's status page. Only major incidents or ones lasting 30 minutes or more are counted. Fewer records do not prove better uptime.

Incidents per week, last 13 weeks. Most in one week: 1.

Aug 3, 2026Last 7 days

Oct 23GitHub reported degraded Grok 4.6 and Grok 4.7 in Copilot42 min

Writing habits

dashes used like commas, per 1,000 words: 4.62
uses of "delve": None in 36 replies
words from our AI cliché list, per 1,000 words: None in 36 replies
"not X but Y" phrases per 1,000 words: 0.09
replies with headings or bullet lists: 19 of 36
typical reply length in words: 294 words
says only "OK" when told to: Yes

How we measured this

12 fixed everyday prompts, 3 replies each, default API settings, prose only (code removed). Method version 3. Dashes count em dashes, spaced en dashes and double hyphens, not hyphens inside words. Measured Oct 24, 2026.