BenchLeader

Grok 4.6 vs Grok 4.7

Full profiles: Grok 4.6 (SpaceXAI) and Grok 4.7 (SpaceXAI): every benchmark, price and rank.

Updated · Data from 12 sources

As of 16 Nov 2026, Grok 4.6 and Grok 4.7 are neck and neck on quality (60.6 vs 57.4 on the BenchLeader Index, inside the ±4.3 margin), and they cost about the same at $3.00 vs $3.00 per 1M tokens blended. Grok 4.6 leads in every category we measure both in.

Key facts (as of 16 Nov 2026)

  • Grok 4.6 scores 60.6 on the BenchLeader Index and Grok 4.7 scores 57.4 (BenchLeader, 16 Nov 2026).
  • Grok 4.6 costs $3.00 per 1M tokens blended (list price, 16 Nov 2026).
  • Grok 4.7 costs $3.00 per 1M tokens blended (list price, 16 Nov 2026).
  • Of the 54 benchmarks both models have, Grok 4.7 leads 31 (BenchLeader, 16 Nov 2026).
  • On Arena WebDev, Grok 4.6 is rated 1618 and Grok 4.7 1641 (LMArena, as published by 16 Nov 2026).

Which to pick for…

The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.

Use casePickWhyConfidence
CodingEither61 vs 61 category score; Arena WebDev 1618 vs 1641Either
Agents & tool useGrok 4.659 vs 57 category score; Terminal-Bench 2.1 (Vals) 78.3% vs 73.4%Medium
MathsGrok 4.663 vs 56 category score; Arena Maths 1445 vs 1447High
ReasoningGrok 4.664 vs 56 category score; GPQA Diamond 94.0% vs 92.7%High
KnowledgeEither60 vs 60 category score; SimpleQA Verified 49.3% vs 56.0%Either
Long documentsGrok 4.665 vs 63 category score; AA-LCR v1.1 80.3% vs 76.7%Medium
Images and other mediaGrok 4.660 vs 56 category score; Arena Vision 1264 vs 1230Medium
Chat quality (human preference)Grok 4.664 vs 63 on Arena votesMedium
Lowest costEither$3.00 vs $3.00 per 1M blended; agentic coding session $0.144 vs $0.144–
Fastest first answerGrok 4.628.6 s (Grok 4.6 (high)) vs 55.9 s (Grok 4.7 (xhigh)) to the first answer token–
High volume on a budgetNeither flagshipA smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs–

Key differences

  • Grok 4.6 (high) and Grok 4.7 (xhigh) are neck and neck on quality (60.6 vs 57.4, within the ±4.3 margin).
  • Grok 4.6 (high) is stronger in agents & tools, coding, human preference, long context, maths, multimodal, reasoning.
  • Grok 4.7 (xhigh) is stronger in knowledge.
  • They cost about the same ($3.00 per 1M blended).

At a glance

SpecGrok 4.6Grok 4.7
DeveloperSpaceXAISpaceXAI
Released11 Aug 202621 Sep 2026
Context window (tokens)500,000500,000
Compared atGrok 4.6 (high)Grok 4.7 (xhigh)
Open weightsNoNo
Price in / out / cached input per 1M$2.00 / $6.00 / $0.500$2.00 / $6.00 / $0.500

Category by category

Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.

CategoryGrok 4.6 (high)Grok 4.7 (xhigh)Leader
Reasoning63.555.6Grok 4.6
Coding61.160.9level
Agents & tools59.256.9Grok 4.6
Maths63.455.8Grok 4.6
Knowledge59.559.9level
Human preference64.162.5Grok 4.6
Multimodal60.055.5Grok 4.6
Long context64.862.9Grok 4.6
Composite–52.8not comparable

On one chart

Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.

2550 avg75ReasoningCodingAgents & toolsMathsKnowledgeLong contextMultimodalHumanpreferenceGrok 4.6, Reasoning: 63.5Grok 4.6, Coding: 61.1Grok 4.6, Agents & tools: 59.2Grok 4.6, Maths: 63.4Grok 4.6, Knowledge: 59.5Grok 4.6, Long context: 64.8Grok 4.6, Multimodal: 60.0Grok 4.6, Human preference: 64.1Grok 4.7, Reasoning: 55.6Grok 4.7, Coding: 60.9Grok 4.7, Agents & tools: 56.9Grok 4.7, Maths: 55.8Grok 4.7, Knowledge: 59.9Grok 4.7, Long context: 62.9Grok 4.7, Multimodal: 55.5Grok 4.7, Human preference: 62.5

Not drawn, measured for fewer than two of these models: Instruction following.

Model, click to add or removeIndex

Where sources disagree

Across the 54 benchmarks both have, Grok 4.6 leads 23, Grok 4.7 leads 31. Independent sources split: Grok 4.6 leads on Arena Text; Grok 4.7 leads on AA Intelligence Index v5.0.1.

Overall indexGrok 4.6 (high)Grok 4.7 (xhigh)Leader
BenchLeader Index60.657.4neck and neck
AA Intelligence Index v5.0.131.0†35.5Grok 4.7
Arena Text14571443Grok 4.6

Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).

Price, speed and latency

List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.

MeasureGrok 4.6 (high)Grok 4.7 (xhigh)
Blended price per 1M tokens$3.00$3.00
Input / output per 1M$2.00 / $6.00$2.00 / $6.00
Cached input per 1M$0.500$0.500
Price per task: chat reply$0.0026$0.0026
Price per task: summarize a 30-page report$0.028$0.028
Price per task: code edit$0.021$0.021
Price per task: agentic coding session$0.144$0.144
Price per task: structured extraction$0.0052$0.0052
Agentic coding session, input mostly cached$0.076$0.076
Cost per benchmark task (AA Index)–$7.77
Output tokens per task (AA Index)–105k
Of which reasoning–69k
Time per task (AA Index)–26 min
Output speed73 tok/s68 tok/s
Time to first answer28.6 s55.9 s
Time to a complete answer, last measured 13 Nov 202662.7 s145.5 s

Each model is compared at its best-ranked reasoning-effort setting. Grok 4.6 (high) scores 60.6 and takes 29 s to its first answer token; Grok 4.6 (medium) scores 59.4 in 19 s, Grok 4.6 (xhigh) scores 56.1 in 37 s. Grok 4.7 (xhigh) scores 57.4 and takes 56 s to its first answer token; Grok 4.7 (high) scores 56.9 in 35 s, Grok 4.7 (low) scores 55.4 in 10 s. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.

Every benchmark

167 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.

MetricPublisherGrok 4.6 (high)Grok 4.7 (xhigh)
BenchLeader IndexBenchLeader60.657.4
Agents & tools scoreBenchLeader59.256.9
Coding scoreBenchLeader61.160.9
Human preference scoreBenchLeader64.162.5
Knowledge scoreBenchLeader59.559.9
Long context scoreBenchLeader64.862.9
Maths scoreBenchLeader63.455.8
Multimodal scoreBenchLeader60.055.5
Reasoning scoreBenchLeader63.555.6
Composite scoreBenchLeader–52.8
Blended price $/MBenchLeader$3.00$3.00
Output speedBenchLeader73 tok/s68 tok/s
Time to first answerBenchLeader28.6 s55.9 s
Context windowBenchLeader500,000 tokens500,000 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub94.0%92.7%
SimpleBenchSimpleBench75.9%†–
Arena Hard PromptsArena (LMArena)14811466
LiveBench 2.0 ReasoningLiveBench91.2%†82.7%
LiveBench 1.0 ReasoningLiveBench90.5%†82.7%
GPQA Diamond (AA)Artificial Analysis95.0%–
Humanity's Last Exam (AA)Artificial Analysis42.9%43.1%
HLE-Diamond (AA)Artificial Analysis21.5%†23.2%
GPQA Diamond (Vals)Vals AI94.7%–
ARC-AGI-1ARC Prize87.0%89.5%
ARC-AGI-2ARC Prize65.1%61.4%
ARC-AGI-3ARC Prize2.1%†7.0%
CritPtArtificial Analysis17.1%17.7%
CritPt v2 hard subset (AA)Artificial Analysis40.0%†33.5%
MysteryMechanismVals AI30.6%25.2%
HLE-Diamond (Scale)Scale AI SEAL–23.4%†
Arena ExpertArena (LMArena)14961478
Chess PuzzlesEpoch AI Benchmarking Hub33.0%41.0%
Mystery Game PuzzlesEpoch AI Benchmarking Hub34.0%†29.0%
LMCAEpoch AI Benchmarking Hub48.5%†49.4%
DTBenchEpoch AI Benchmarking Hub97.3%†96.0%
ForecastBenchForecastBench65.4%†–
Coding
SciCodeSciCode56.5%57.4%
WeirdMLWeirdML67.3%–
FrontierCodeCognition48.0%†47.6%†
Show all 167 benchmarks
MetricPublisherGrok 4.6 (high)Grok 4.7 (xhigh)
Coding
Arena CodingArena (LMArena)15081493
Arena WebDevArena (LMArena)16181641
LiveBench 2.0 CodingLiveBench76.7%†73.1%
LiveBench 1.0 CodingLiveBench76.8%†77.2%
LiveCodeBenchVals AI88.2%–
IOIVals AI47.6%57.7%
SWE-bench (Vals)Vals AI95.6%–
Code MigrationVals AI44.6%44.8%
Code Migration v1.1Vals AI–53.2%
ProgramBenchVals AI1.0%0.5%
SWE-bench MultimodalVals AI–42.9%
Vibe Code Bench 1-100Vals AI14.8%–
Vibe Code Bench v1.1Vals AI76.2%86.2%
Vibe Code Bench 1-100 v1.1Vals AI–26.4%
SWE Atlas v1.1Scale AI SEAL–62.1%†
SWE Atlas v1.1: Codebase QnAScale AI SEAL–73.4%†
SWE Atlas v1.1: Test WritingScale AI SEAL–64.4%†
SWE Atlas v1.1: RefactoringScale AI SEAL–48.6%†
MirrorCodeEpoch AI Benchmarking Hub41.1%†54.4%
DeepSWE v1.1Epoch AI Benchmarking Hub65.2%–
CursorBenchCursor40.4%46.3%
ALE-BenchEpoch AI Benchmarking Hub1508.1†1545.8
FrontierSWEEpoch AI Benchmarking Hub25.3%†29.5%
WeirdML v3WeirdML–8.4%
Terminal-Bench 4.0 (AA)Artificial Analysis21.2%25.8%
Terminal-Bench 2.1 (AA)Artificial Analysis88.4%–
Startrise FrontendStartrise72.1%†75.9%†
GameBench 2 OverallDuelLab GameBench 259.4†63.6†
GameBench 2 by reasoning settingDuelLab GameBench 268.7†71.2
Design Arena WebsiteDesign Arena1277†1244†
FrontierCode 1.1 (AA)Artificial Analysis–33.8%
FrontierCode 2.0 (AA)Artificial Analysis19.8%†28.7%
Agents & tools
Terminal-BenchTerminal-Bench19.7%38.8%
APEX-AgentsMercor65.3%†54.6%†
Arena AgentArena (LMArena)-0.2†2.6
LiveBench Long AgenticLiveBench18.7%†45.3%
LiveBench BrowserLiveBench49.4%†49.3%
LiveBench Agentic CodingLiveBench57.0%†54.0%
Terminal-Bench 2.1 (Vals)Vals AI78.3%73.4%
GDPval-AA v2.1, normalised scoreArtificial Analysis–60.7%
GDPval-AA v2.1Artificial Analysis16271717
τ³-Banking (AA)Artificial Analysis50.7%–
ITBench SRE (AA)Artificial Analysis–42.1%
Analyst Agent (AA)Artificial Analysis41.3%52.5%
CyberBenchVals AI67.7%69.5%
Finance Agent v2Vals AI53.7%52.3%
Finance Agent v2.1Vals AI–60.6%
Harvey's Legal Agent Benchmark v1.1Vals AI–18.3%
Harvey's Legal Agent Benchmark v1Vals AI15.8%12.5%
Legal Research BenchVals AI48.1%47.1%
Robo-benchVals AI–0.2%
SkillsBenchVals AI55.8%–
SREBenchVals AI–3.8%
Tax Agent BenchVals AI33.7%35.8%
Terminal-Bench 4.0 (Vals)Vals AI17.2%28.8%
Terminal-Bench ScienceVals AI11.4%10.0%
Time Horizon Index: KSPVals AI5.8%–
Arena Agent Task CompletionArena (LMArena)-2.6†5.1
Vending-Bench 2Andon Labs9047.0†10536.8†
AutomationBenchArtificial Analysis66.7%65.6%
GDP.pdf (AA)Artificial Analysis17.0%20.0%
Harvey LAB-AA v1.1Harvey–9.4%
EnterpriseOps-GymArtificial Analysis48.3%–
AA-Briefcase v1.1Artificial Analysis15261636
GDPxGDPx by SemiAnalysis–3.9%
Design Arena AgenticDesign Arena1239†1277†
Terminal-Bench-Science 0.1 (AA)Artificial Analysis6.2%11.4%
CWE-Bench (AA)Artificial Analysis–68.3%
DeepSec-Bench (AA)Artificial Analysis–26.9%
CyberGym end-to-end (AA)Artificial Analysis–74.0%
MKA (AA)Artificial Analysis–40.1%
FACTS SearchGoogle87.9%†85.6%†
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub66.0%†53.0%
FrontierMath Tier 4Epoch AI Benchmarking Hub31.7%†17.1%
OTIS Mock AIMEEpoch AI Benchmarking Hub97.8%98.1%
ProofBenchVals AI51.0%†26.0%†
LiveBench MathematicsLiveBench92.6%†95.7%
Arena MathsArena (LMArena)14451447
FrontierMath ErdősEpoch AI Benchmarking Hub–0.0%
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub49.3%56.0%
LiveBench 2.0 Data AnalysisLiveBench75.3%†76.9%
LiveBench ResearchLiveBench88.8%†89.5%
LiveBench 1.0 Data AnalysisLiveBench73.9%†76.9%
AA-Omniscience v1.1Artificial Analysis24.9†31.2
MMLU-ProVals AI89.4%–
LegalBenchVals AI86.3%84.4%
CorpFinVals AI66.2%–
TaxEvalVals AI71.1%–
PRBench LegalScale AI SEAL–47.6%
BioMysteryBenchVals AI72.2%69.3%
Excel Modeling BenchmarkVals AI62.7%67.0%
Excel Modeling Benchmark v1.1Vals AI–71.5%
MedCodeVals AI44.7%49.5%
MedScribeVals AI86.5%89.4%
Public Benefits Bench v1.1Vals AI66.8%65.6%
Access to Justice BenchmarkVals AI71.3%73.7%
Financial Coaching BenchVals AI–74.1%
Immigration BenchVals AI66.0%68.0%
EBR-benchEpoch AI Benchmarking Hub30.5%†–
GDP.pdfEpoch AI Benchmarking Hub16.0%22.8%
AA-Omniscience v1.1: accuracyArtificial Analysis46.2%†42.7%
AA-Omniscience v1.1: non-hallucinationArtificial Analysis60.3%†80.0%
HealthBench Professional (AA)OpenAI–55.8%
FACTS ParametricGoogle44.5%†54.3%†
Instruction following
LiveBench LanguageLiveBench83.7%†80.1%
LiveBench Instruction FollowingLiveBench71.9%†75.3%
Arena Instruction FollowingArena (LMArena)14511435
Human preference
Arena TextArena (LMArena)14571443
EQ-Bench Creative Writing v3EQ-Bench–2012†
Arena Creative WritingArena (LMArena)14421421
Arena Multi-turnArena (LMArena)14491435
Arena Longer QueriesArena (LMArena)14761454
Multimodal
Arena VisionArena (LMArena)12641230
LiveBench DesignLiveBench76.3%†75.0%
MMMU-ProArtificial Analysis–78.2%†
MortgageTaxVals AI64.2%–
SAGEVals AI–88.7%
GeoBenchGeoBench–61.0%†
Blueprint-Bench 2Epoch AI Benchmarking Hub33.2%†32.5%†
Furniture AssemblyEpoch AI Benchmarking Hub40.0%†20.8%
VIALS-AAHandshake AI–37.9%
Code2Video-AAArtificial Analysis–1127
AA-3DArtificial Analysis–992
FACTS MultimodalGoogle39.3%†40.9%†
Long context
AA-LCR v1.1Artificial Analysis80.3%76.7%
Arena DocumentArena (LMArena)1450–
MLCRArtificial Analysis15.6%20.0%
FACTS GroundingGoogle66.5%†68.6%†
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub157.8†156.2†
LiveBench 2.0LiveBench63.5%†67.9%
LiveBench 1.0LiveBench78.0%†77.4%
AA Intelligence Index v5.0.1Artificial Analysis31.0†35.5
AA Intelligence Index v4.3.2Artificial Analysis44.346.5
AA Progress Index v1Artificial Analysis321.5322.7
Vals IndexVals AI–58.6
Vals Finance IndexVals AI–58.8%
Vals RSI IndexVals AI25.5%†20.2%
Vals Science IndexVals AI38.1%34.8%
Vals Router IndexVals AI–58.6%†
Vals Legal IndexVals AI–72.7%
AA Cyber IndexArtificial Analysis–56.4
Safety & honesty
DistressBenchScale AI SEAL–52.4%†

Data as of 2026-11-16. Best configuration of each model; every score links to its source on the model pages.

DuelLab GameBench 2, release gb2-20261111-7b598bb13de4 (https://benchmarks.duellab.org/releases/gb2-20261111-7b598bb13de4/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.

What changed

  • Grok 4.7 (xhigh): BenchLeader Index 58.0 → 57.4.
  • Grok 4.7 (xhigh): 8 new results (LiveBench 2.0, LiveBench Long Agentic, LiveBench 2.0 Reasoning…).
  • Grok 4.6 (high): 1 new result (GDPval-AA v2.1).
  • Grok 4.6 (high): BenchLeader Index 60.9 → 60.4.
  • Grok 4.7 (xhigh): 1 new result (MirrorCode).
  • Grok 4.7 (xhigh): 1 new result (ALE-Bench).
  • Grok 4.6 (high): 2 score revisions.
  • Grok 4.7 (xhigh): 2 score revisions.
  • Grok 4.6 (high): 1 score revision.
  • Grok 4.7 (xhigh): 2 score revisions.

Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs

Grok 4.6 vs Grok 4.7: questions

Is Grok 4.6 better than Grok 4.7?
Grok 4.6 (high) and Grok 4.7 (xhigh) are neck and neck on quality (60.6 vs 57.4, within the ±4.3 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 16 Nov 2026, so check the category scores for your use.
Grok 4.6 vs Grok 4.7: which is cheaper, and will that change?
They cost the same: $3.00 per million tokens blended. No price change has been announced for either.
Is Grok 4.6 or Grok 4.7 better for coding?
They are level in coding (61 vs 61 on the category index). On Arena WebDev, the top shared coding benchmark, it is 1618 for Grok 4.6 against 1641 for Grok 4.7.
Which is faster, Grok 4.6 or Grok 4.7?
Grok 4.6 streams faster (73 vs 68 output tokens per second, Artificial Analysis 7-day median).
Is Grok 4.6 or Grok 4.7 better for agents and tool use?
Grok 4.6 scores higher in agentic tasks (59 vs 57 on the category index, where 50 is average). On Terminal-Bench, the top shared agents & tools benchmark, it is 19.7% for Grok 4.6 against 38.8% for Grok 4.7.
More questions (4)
Which is better for long documents?
Grok 4.6 scores higher on long-context benchmarks (65 vs 63). Both accept 500,000 tokens of context.
Which has lower latency?
Grok 4.6 answers first: 28.6 s against 55.9 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
Do independent benchmarks agree on Grok 4.6 vs Grok 4.7?
Across the 54 benchmarks both have, Grok 4.6 leads 23, Grok 4.7 leads 31. Independent sources split: Grok 4.6 leads on Arena Text; Grok 4.7 leads on AA Intelligence Index v5.0.1.
When was Grok 4.7 released?
Grok 4.7 was released on 21 Sep 2026.