Claude Opus 4.7 vs GPT-5.4 — Who Actually Wins? Real Benchmark Data (2026)

head to head

Claude Opus 4.7 vs GPT-5.4

14 shared benchmarks

9

wins

0

ties

5

wins

75.8%

ARC-AGI 2

73.3%

90.9%

BigLaw Bench

91.0%

79.3%

BrowseComp

82.7%

73.8%

CyberGym

79.0%

43.8%

FrontierMath

47.6%

94.2%

GPQA Diamond

92.8%

46.9%

Humanity's Last Exam

36.24%

75.2%

MMMU Pro

81.2%

59.2%

MRCR v2

36.6%

78.0%

OSWorld

75.0%

78.0%

OSWorld-Verified

75.0%

64.3%

SWE-bench Pro

57.7%

87.6%

SWE-bench Verified

77.2%

77.0%

Terminal-Bench 2.0

75.1%