Compare / Claude Sonnet 4.6 vs Grok 4.3

Claude Sonnet 4.6 vs Grok 4.3

26 results from 6 sources that measured both models. Each row is in the source's own units; there is no overall winner.

Claude Sonnet 4.6
Anthropic · profile
Grok 4.3
xAI · profile
Results compared
26
Shared sources
6
BenchmarkMetricClaude Sonnet 4.6Grok 4.3
Arena (formerly LMArena)reportedOverallArena rating 1,2011,193
Arena (formerly LMArena)reportedOverallArena rating 1,4611,428
Arena (formerly LMArena)reportedBusiness, management and financeArena rating 1,4821,440
Arena (formerly LMArena)reportedCreative writingArena rating 1,4491,423
Arena (formerly LMArena)reportedExpert promptsArena rating 1,5061,452
Arena (formerly LMArena)reportedInstruction followingArena rating 1,4751,415
Arena (formerly LMArena)reportedOverallArena rating 1,4721,442
Arena (formerly LMArena)reportedWriting, literature and languageArena rating 1,4551,424
Artificial AnalysisreportedArtificial Analysis Coding Indexindex score 63.042.2
Artificial AnalysisreportedArtificial Analysis Intelligence Indexindex score 30.124.9
Artificial AnalysisreportedGPQA Diamond% of questions 87.5%90.1%
Artificial AnalysisreportedHumanity's Last Exam% of questions 33.6%37.2%
Artificial AnalysisreportedIFBench% of instructions 56.6%83.3%
Artificial AnalysisreportedLong-context reasoning (AA-LCR)% of questions 80.0%75.0%
Artificial AnalysisreportedSciCode% of problems 50.1%48.3%
Artificial AnalysisreportedTerminal-Bench 2.1% of tasks 71.2%39.7%
Artificial AnalysisreportedTerminal-Bench 4.0% of tasks 3.0%0.0%
Artificial AnalysisreportedTerminal-Bench Hard% of tasks 53.0%37.9%
Artificial AnalysisreportedΤ-bench banking% of tasks 34.4%12.4%
Artificial AnalysisreportedΤ²-bench telecom% of tasks 79.5%97.7%
GDP.pdfreportedProfessional document tasks% of rubric 18.0%8.0%
SimpleQA Verified (Epoch AI)reportedCorrect answers% of questions 35.5%33.2%
UGI LeaderboardreportedRequested-length error% off the requested word count 2.0%21.0%
UGI LeaderboardreportedStyle adherencescore from 0 to 1 0.410.34
UGI LeaderboardreportedWriting scorescore out of 100 64.557.7
Vending-Bench 2reportedMoney after a yearUS dollars $7,204.14$35.26

Bold green marks the better value on that metric. Where a source reports ranges that overlap, the difference may not be meaningful; see the benchmark page for ranges.