Inference
EN
xAI

Grok 4.7

Grok 4.7 is a reasoning model from xAI. 55 benchmarks count toward its index, in 6 categories.

  • Closed weights
  • Reasoning
  • Input: text · image · file
  • Released 2026-09
RankedEnough results for an index.
IndexAll categories, out of 100.69.1±3.7Median 51.0
CoverageIndex weight that has results.85%
SpeedOutput tokens per second.51/sMedian 67/s
First tokenSeconds to the first output token.5.1 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$2.00batch $1.60Median $0.60
Output / 1MUS dollars per 1M output tokens.$6.00batch $4.80 Batch price per 1M output tokens.Median $2.55
ContextMaximum tokens in one request.500KMedian 262K
EloLMArena rating and rank.1398±5 #160Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
x-ai/grok-4.7
Max outputMaximum output tokens per request.
450K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.50
WeightsLink to the published weights.
—
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
4 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
Amazon BedrockCloud$2.00$6.00500K
Google Vertex AICloud$2.00$6.00524K
OpenRouterRouter$2.00$6.00500K
xAIMaker$2.00$6.00500K

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.22 counted71.916 / 218
CodingCode writing and repair.11 counted69.920 / 209
ReasoningLogic problems and puzzles.10 counted64.546 / 227
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.6 counted66.535 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted78.9—
MathMath problems.5 counted61.841 / 168

Results

55 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
Artificial Analysis CWE-Bench-AAAgentic68.3%1 / 19
GDPval-AA normalizedAgentic60.7%8 / 133
LiveBench ResearchAgentic86.4%5 / 52
Artificial Analysis Cyber IndexAgentic56.4%2 / 19
Vibe Code Bench v1.1Coding86.2% ±2.213 / 123
GPQA diamond · Epoch AIKnowledge92.7% ±1.625 / 217
Artificial Analysis Humanity's Last ExamKnowledge43.1%32 / 222
LiveBench Instruction FollowingInstruction75.3%10 / 66
OTIS Mock AIME 2024-2025Math98.1% ±1.431 / 200
CWE-bench v1Agentic68.0%2 / 12
Artificial Analysis SciCodeCoding57.4%19 / 114
Critical Physics TasksReasoning17.7%37 / 214
Agent Arena command recoveryAgentic7.29 / 50
LiveBench MathematicsMath95.7%12 / 66
LiveBench Long AgenticAgentic50.1%11 / 51
Artificial Analysis DeepsecBench-AAAgentic26.9%5 / 21
Code MigrationCoding44.8% ±4.219 / 77
SimpleQA VerifiedKnowledge56.0% ±1.623 / 91
Artificial Analysis Omniscience AccuracyKnowledge42.7%42 / 164
Agent Arena task outcomeAgentic5.114 / 54
Artificial Analysis AnalystAgentAgentic52.5%7 / 26
ARC-AGI-2 (semi-private)Reasoning61.4%36 / 131
Terminal-Bench 2.1 · Vals AIAgentic73.4% ±1.522 / 75
ARC-AGI-1 (semi-private)Reasoning89.5%37 / 126
Artificial Analysis CyberGym-E2E-AAAgenticAbove fitFar above this model's index.74.0%6 / 19
MazeBench: visual spatial reasoning in a 3D open worldAgentic1.0%16 / 48
DeepSWEAgentic71.0%21 / 59
MirrorCodeCoding54.4% ±11.414 / 39
FrontierCode 2.0Coding28.6%10 / 27
LiveBench Data AnalysisReasoning76.9%30 / 80
Hermes Index: agentic capability and cost per task inside Hermes AgentAgentic39.3%6 / 16
Chess PuzzlesReasoning41.0% ±4.934 / 90
LiveBench DesignAgentic78.0%20 / 52
Artificial Analysis Long Context ReasoningReasoning76.7%84 / 217
Artificial Analysis AutomationBenchAgentic65.6%9 / 23
IOICoding57.7% ±2.018 / 45
Mystery Game PuzzlesReasoning29.0% ±4.632 / 80
Terminal-Bench 4.0Agentic28.8% ±2.325 / 61
LiveBench LanguageKnowledge80.1%28 / 66
Terminal-Bench 4.1.0Agentic38.8% ±1.810 / 23
Terminal-Bench 4.0.0Agentic37.6% ±1.810 / 23
LiveBench Coding 2.0Coding84.5%24 / 52
ARC-AGI-3 (semi-private)Reasoning1.8%12 / 26
LiveBench Agentic CodingAgentic54.0%33 / 66
cursorBench40Coding46.3%11 / 21
CWE-bench v2Agentic54.7%6 / 11
Bug Hunt BenchCoding29.0%14 / 25
LiveBench CodingCoding77.2%37 / 66
LiveBench Browser UseAgentic60.0%28 / 49
FrontierMath-Tiers-1-3-v2-PrivateMath53.0% ±3.053 / 90
FrontierSWE v2Coding29.5%13 / 22
Multimodal Knowledge Agent BenchmarkAgentic40.1%12 / 19
LiveBench ReasoningReasoning82.7%51 / 80
Agent Arena steerabilityAgentic−2.335 / 54
Artificial Analysis ITBench-AAAgentic42.1%16 / 24
Medical Long Context Reasoning (MLCR-AA)Reasoning20.0%15 / 22
FrontierMath-Tier-4-v2-PrivateMath17.1% ±5.952 / 72
Artificial Analysis Terminal-Bench v4.0Agentic25.8%16 / 22
ProofBench v1.1Math26.0% ±4.440 / 54
Artificial Analysis GDP.pdfAgentic20.0%17 / 22
HealthBench ProfessionalKnowledge56.7%15 / 18
Furniture AssemblyReasoning20.8% ±5.236 / 37
MedScribeHealthcare89.4% ±1.97 / 109
Legal Research BenchLegal47.1% ±3.516 / 90
Tax Agent BenchFinance35.8% ±3.516 / 82
MedCodeHealthcare49.6% ±2.221 / 106
HLAB-Vals-1.1Legal18.3% ±3.27 / 35
LegalBenchLegal84.4% ±0.533 / 149
SAGEEducation88.7% ±1.87 / 29
EMBFinance67.0% ±3.020 / 75
ARC-AGI-1 (public eval)Reasoning95.1%32 / 116
ProgramBenchCoding0.5% ±0.519 / 65
ARC-AGI-2 (public eval)Reasoning60.0%39 / 122
Terminal-Bench-ScienceScience10.0% ±3.616 / 44
Public Benefits Bench v1.1Public benefits65.6% ±1.221 / 50
Finance Agent (v2)Finance52.3% ±0.436 / 79
CyberBench v1.1Other69.5% ±5.724 / 48
BioMysteryBenchScience69.3% ±1.315 / 30
SRE BenchOther3.8% ±1.219 / 36
FrontierMath-ErdosMath0.0%8 / 15
MysteryMechanismScience25.2% ±2.917 / 30
Hermes Bench: 150 agent tasks graded on the files and state the agent leaves behindAgentic74.4%—
EEBench V1 core corpusCoding64.0%—
Artificial Analysis Harvey LAB-AA v1.0Agentic19.6%—

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseEpoch AICC BYLiveBenchNo licence statedARC Prize FoundationTerms restrict reuseVals AINo licence statedTerminal-BenchNo licence statedLMArenaCC BY 4.0models.devMIT

History

5 releases
Index69.5 → 69.1
23 Oct 202612 Nov 2026
Output / 1M$6.00 → $6.00
23 Oct 202612 Nov 2026

Questions

How much does Grok 4.7 cost?
$2.00 per 1M input tokens, $6.00 per 1M output tokens.
How good is Grok 4.7?
Its index is 69.1 out of 100. That is place 24 of 283 ranked models.
What is Grok 4.7 best at?
Agentic: 71.9 out of 100, place 16 of 218.
How fast is Grok 4.7?
51 output tokens per second, 5.1 s to the first token.
What is the context window of Grok 4.7?
500K tokens.
Is Grok 4.7 open source?
No. Its weights are not published.
When was Grok 4.7 released?
September 2026.

Badge

Grok 4.7 on Inference360: place 24 of 283, index 69.1
HTML
<a href="https://inference360.ai/models/grok-4-7"><img src="https://inference360.ai/badges/grok-4-7.svg" alt="Grok 4.7 on Inference360: place 24 of 283, index 69.1" height="20"></a>
Markdown
[![Grok 4.7 on Inference360: place 24 of 283, index 69.1](https://inference360.ai/badges/grok-4-7.svg)](https://inference360.ai/models/grok-4-7)

Compare with

GPT-5.569.1MiniMax M3.1 Flash69.3GPT-5.6 Terra68.9DeepSeek V4.1 Flash 101768.5Gemini 3.7 Flash68.5

Lower price

Ling 3.1 Flash67.2 · $0.135/1MDeepSeek V4.1 Flash 101768.5 · $0.40/1MClaude Haiku 5.566.3 · $0.50/1M

More from xAI

All models
Grok 4.667.8Grok LatestUnrankedGrok 4.563.7Grok Build 0.152.7Grok 4.353.3