The Model Index
Up to four models

Kimi K3 vs Claude Sonnet 5

On the index (GPT-4 = 100), Kimi K3 scores 168 and Claude Sonnet 5 164. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.

Next to Kimi K3
Presets
Kimi K3
168152–183
WeightsOpen
ReleasedJul 2026
Price in / out$0.19 / $16
Context1.05M
Core tests taken6 of 9
Claude Sonnet 5
164148–180
WeightsClosed
ReleasedJun 2026
Price in / out$2.0 / $10
Context1M
Core tests taken5 of 9

Kimi K3 and Claude Sonnet 5 are too close to call (3 points apart, within the uncertainty), ahead on 3 of 5 shared core tests, at a similar price.

Test by testthird-party results

Kimi K3Claude Sonnet 5
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Epoch AI)GPQA Diamond (Epoch AI)LiveBench Data AnalysisLiveBench Data AnalysisLiveBench IFLiveBench IFLiveBench LanguageLiveBench LanguageLiveBench ReasoningLiveBench ReasoningGPQA Diamond (Vals AI)GPQA Diamond (Vals AI)MMLU ProMMLU ProMMMU ProMMMU ProSimpleBenchSimpleBenchSimpleQA VerifiedSimpleQA VerifiedMATHSFrontierMath (tiers 1-3)FrontierMath (tiers 1-3)AIME-style maths (OTIS mock)AIME-style maths (OTIS mock)LiveBench MathematicsLiveBench MathematicsProofBench v1.1ProofBench v1.1Chess puzzlesChess puzzlesFrontierMath (tier 4)FrontierMath (tier 4)Surface Evolver BenchSurface Evolver BenchCODINGTerminal-Bench 4.0Terminal-Bench 4.0LiveBench Agentic CodingLiveBench Agentic CodingLiveBench CodingLiveBench CodingCode MigrationCode MigrationCyberBench v1.1CyberBench v1.1IOIIOILiveCodeBenchLiveCodeBenchProgramBenchProgramBenchSWE-benchSWE-benchVibe Code Bench 1-100Vibe Code Bench 1-100Vibe Code Bench v1.1Vibe Code Bench v1.1DeepSWEDeepSWEMirrorCodeMirrorCodeWeirdMLWeirdMLAGENTSAPEX-AgentsAPEX-AgentsTerminal-Bench-ScienceTerminal-Bench-ScienceDTBenchDTBenchLMCALMCAMystery game puzzlesMystery game puzzlesPROFESSIONAL WORKFinance Agent (v2)Finance Agent (v2)CorpFin v2CorpFin v2EMBEMBLegalBenchLegalBenchLegal Research BenchLegal Research BenchMedCodeMedCodeMedScribeMedScribeMortgageTaxMortgageTaxPublic Benefits Bench v1.1Public Benefits Bench v1.1Tax Agent BenchTax Agent BenchTaxEval v2TaxEval v2
Every result as a table
TestKimi K3Claude Sonnet 5
Reasoning & knowledge
GPQA Diamond (Epoch AI)93.1%90.5%
LiveBench Data Analysis78.7%71.7%
LiveBench IF71.4%63.9%
LiveBench Language85.5%75.0%
LiveBench Reasoning90.7%88.7%
GPQA Diamond (Vals AI)92.9%88.9%
MMLU Pro88.0%87.5%
MMMU Pro88.1%83.0%
SimpleBench60.7%60.6%
SimpleQA Verified50.6%33.7%
Maths
FrontierMath (tiers 1-3)72.2%65.6%
AIME-style maths (OTIS mock)97.2%94.7%
LiveBench Mathematics84.4%92.9%
ProofBench v1.187.0%77.0%
Chess puzzles37.0%45.0%
FrontierMath (tier 4)39.0%29.3%
Surface Evolver Bench95.0%60.0%
Coding
Terminal-Bench 4.017.2%10.6%
LiveBench Agentic Coding62.2%59.4%
LiveBench Coding81.5%80.7%
Code Migration16.1%44.4%
CyberBench v1.175.2%70.3%
IOI48.9%45.0%
LiveCodeBench87.2%82.4%
ProgramBench2.0%0.0%
SWE-bench93.4%79.6%
Vibe Code Bench 1-10018.2%13.8%
Vibe Code Bench v1.185.0%81.3%
DeepSWE68.5%53.8%
MirrorCode54.4%63.3%
WeirdML82.6%69.0%
Agents
APEX-Agents47.8%54.5%
Terminal-Bench-Science1.4%5.7%
DTBench91.2%92.5%
LMCA52.7%50.0%
Mystery game puzzles26.0%35.0%
Professional work
Finance Agent (v2)53.1%53.9%
CorpFin v271.6%67.0%
EMB66.7%66.3%
LegalBench86.2%83.9%
Legal Research Bench46.2%41.8%
MedCode49.4%47.5%
MedScribe88.0%76.0%
MortgageTax66.3%70.0%
Public Benefits Bench v1.168.2%66.0%
Tax Agent Bench33.4%29.4%
TaxEval v275.7%75.6%

Tests at least two of these models have taken. Bold rows are the core tests.

This page as a chartDownload PNG