NOUMENOpen app
NOUMEN

xAI · Frontier

Grok 4.7

Added Sep 21, 2026

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

Grok 4.7 is SpaceXAI's flagship reasoning model for coding, long-running agentic tasks, and knowledge work. It supports text, image, and file inputs with tool calling and structured outputs. Reasoning is always active, and requests above 200k input tokens use higher long-context rates. Content policy rejections can still be charged: SpaceXAI may pass through a $0.05 moderation-failure fee, or a $0.055 usage-guidelines violation fee, depending on which rejection upstream returns.

  • Reasoning
  • Tools
  • Vision
  • Structured output
Input
Text, Images
Output
Text
Reasoning effort
low · medium · high · xhigh
Discovery tier
Frontier

Pricing

USD per 1M tokens
Input
$2.00
Output
$6.00
Cache read
$0.50

Reference prices as published by the model provider. API example

Specifications

Context window
500K
Max output
—
Average output (7 days)
1.3K tokens
Longer than 80% of 473 models

Side by side

Grok 4.7 compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Grok 4.7xAIThis modelGPT 5.6 SolOpenAIRemove
ProviderxAIOpenAI
Discovery tierFrontierFrontier
AddedSep 21, 2026Jul 9, 2026
Limits
Context window500K1.05MBest
Max output—128K
Average output1.3K tokens576 tokens
Pricing, USD per 1M tokens
Input$2.00$2.00
Output$6.00Best$10.00
Cache read$0.50$0.20Best
Cache write—$2.50
Performance
Throughput238.7 tok/sBest72.0 tok/s
Time to first token3.9 sBest13.5 s
Request success (24h)99.9%100.0%Best
Benchmark indexes
Intelligence Index46.347.0Best
Coding Index—77.4
Capabilities
ReasoningYesYes
ToolsYesYes
VisionYesYes
Structured outputYesYes
PDF inputNoYes
Reasoning effortlow, medium, high, xhighnone, low, medium, high, xhigh, max

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • Grok 4.7
  • GPT 5.6 Sol

Intelligence Index

  • Grok 4.746.3
  • GPT 5.6 Sol47.0

Grok 4.7 tested at high reasoning effort

Grok 4.7: Better than 92% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • Grok 4.7: —
    • GPT 5.6 Sol: 77.4

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • Grok 4.7: 63.5%Top
    • GPT 5.6 Sol: 60.1%

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • Grok 4.7: 37.7%Top
    • GPT 5.6 Sol: 30.3%

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • Grok 4.7: 1,630 EloTop
    • GPT 5.6 Sol: 1,479 Elo

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • Grok 4.7: 1,707 EloTop
    • GPT 5.6 Sol: 1,609 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • Grok 4.7: 23.2%
    • GPT 5.6 Sol: 27.2%Top

  • AA-LCRv1.1

    Reasoning across long documents

    • Grok 4.7: 77.0%
    • GPT 5.6 Sol: 84.0%Top

  • MLCR-AA

    Long-context reasoning over medical records

    • Grok 4.7: 18.9%
    • GPT 5.6 Sol: 30.0%Top

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Grok 4.7: 42.3%
    • GPT 5.6 Sol: 49.5%Top

  • GPQA Diamond

    Graduate-level science questions

    • Grok 4.7: —
    • GPT 5.6 Sol: 94.1%

  • IFBench

    Following precise output instructions

    • Grok 4.7: —
    • GPT 5.6 Sol: 72.7%

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • Grok 4.7: 24.7%
    • GPT 5.6 Sol: 39.9%Top

  • SciCode

    Scientific computing in Python

    • Grok 4.7: 57.8%Top
    • GPT 5.6 Sol: 57.1%

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • Grok 4.7: —
    • GPT 5.6 Sol: 65.9%

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • Grok 4.7: —
    • GPT 5.6 Sol: 85.1%

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Grok 4.7: 77.0%
    • GPT 5.6 Sol: 84.0%Top

Scores from Artificial Analysis, updated Oct 26, 2026. Grok 4.7 was tested at high reasoning effort; compare scores measured at the same effort.Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Throughput
238.7 tok/s
Faster than 89% of 257 models
Time to first token
3.9 s
Quicker than 40% of 257 models
Prompt cache hits
18.6%
About $1.72 per 1M input tokens with caching
Request success (24h)
99.9%
Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 92.8%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "x-ai/grok-4.7",
    "reasoning_effort": "medium",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="x-ai/grok-4.7",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

xAI

Other xAI models, starting with the same family.