The Model Index
Up to four models

Claude Opus 4.7 vs Claude Opus 4.6

On the index (GPT-4 = 100), Claude Opus 4.7 scores 167 and Claude Opus 4.6 165. 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 Claude Opus 4.7
Presets
Claude Opus 4.7
167151–183
WeightsClosed
ReleasedApr 2026
Price in / out$5.0 / $25
Context1M
Core tests taken6 of 9
Claude Opus 4.6
165149–181
WeightsClosed
ReleasedFeb 2026
Price in / out$5.0 / $25
Context1M
Core tests taken7 of 9

Claude Opus 4.7 and Claude Opus 4.6 are too close to call (2 points apart, within the uncertainty), ahead on 4 of 5 shared core tests, at a similar price.

Test by testthird-party results

Claude Opus 4.7Claude Opus 4.6
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Epoch AI)GPQA Diamond (Epoch AI)Humanity's Last ExamHumanity's Last ExamLiveBench Data AnalysisLiveBench Data AnalysisLiveBench IFLiveBench IFLiveBench LanguageLiveBench LanguageLiveBench ReasoningLiveBench ReasoningGPQA Diamond (Vals AI)GPQA Diamond (Vals AI)MMLU ProMMLU ProMMMU ProMMMU ProEBR-benchEBR-benchSimpleBenchSimpleBenchSimpleQA VerifiedSimpleQA VerifiedMATHSFrontierMath (tiers 1-3)FrontierMath (tiers 1-3)AIME-style maths (OTIS mock)AIME-style maths (OTIS mock)LiveBench MathematicsLiveBench MathematicsChess puzzlesChess puzzlesFrontierMath (2025 set)FrontierMath (2025 set)FrontierMath tier 4 (2025 set)FrontierMath tier 4 (2025 set)FrontierMath (tier 4)FrontierMath (tier 4)CODINGSWE-bench VerifiedSWE-bench VerifiedLiveBench Agentic CodingLiveBench Agentic CodingLiveBench CodingLiveBench CodingLiveCodeBenchLiveCodeBenchSWE-benchSWE-benchVibe Code Bench v1.1Vibe Code Bench v1.1GSO-BenchGSO-BenchMirrorCodeMirrorCodeWeirdMLWeirdMLAGENTSAPEX-AgentsAPEX-AgentsDTBenchDTBenchFurniture assemblyFurniture assemblyLMCALMCAMystery game puzzlesMystery game puzzlesPROFESSIONAL WORKCorpFin v2CorpFin v2LegalBenchLegalBenchMedCodeMedCodeMedScribeMedScribeMortgageTaxMortgageTaxTaxEval v2TaxEval v2
Every result as a table
TestClaude Opus 4.7Claude Opus 4.6
Reasoning & knowledge
GPQA Diamond (Epoch AI)90.1%90.5%
Humanity's Last Exam36.2%34.4%
LiveBench Data Analysis78.3%69.9%
LiveBench IF66.7%63.3%
LiveBench Language77.9%83.3%
LiveBench Reasoning87.2%88.7%
GPQA Diamond (Vals AI)90.1%89.6%
MMLU Pro89.9%89.1%
MMMU Pro85.5%83.9%
EBR-bench19.1%12.7%
SimpleBench61.7%67.6%
SimpleQA Verified51.7%47.0%
Maths
FrontierMath (tiers 1-3)70.2%66.0%
AIME-style maths (OTIS mock)97.8%94.4%
LiveBench Mathematics92.8%89.3%
Chess puzzles41.0%21.0%
FrontierMath (2025 set)43.8%40.7%
FrontierMath tier 4 (2025 set)22.9%22.9%
FrontierMath (tier 4)31.7%26.8%
Coding
SWE-bench Verified83.5%78.7%
LiveBench Agentic Coding50.7%49.0%
LiveBench Coding82.1%78.2%
LiveCodeBench85.1%84.7%
SWE-bench82.0%78.2%
Vibe Code Bench v1.171.0%57.6%
GSO-Bench44.1%41.2%
MirrorCode43.3%33.3%
WeirdML76.4%78.0%
Agents
APEX-Agents49.2%46.3%
DTBench94.7%91.2%
Furniture assembly33.3%28.3%
LMCA52.2%55.8%
Mystery game puzzles28.0%25.0%
Professional work
CorpFin v266.1%67.0%
LegalBench85.3%85.3%
MedCode54.9%49.1%
MedScribe83.0%86.7%
MortgageTax70.3%68.5%
TaxEval v275.3%76.0%
Task length · METR47 sec

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

This page as a chartDownload PNG