Claude Opus 4.6 vs Gemma 4 31B — Who Actually Wins? Real Benchmark Data (2026)

head to head

Claude Opus 4.6 vs Gemma 4 31B

11 shared benchmarks

9

wins

0

ties

2

wins

100.0%

AIME

89.2%

91.3%

GPQA Diamond

84.3%

40.0%

Humanity's Last Exam

19.5%

76.0%

LiveCodeBench

80.0%

82.0%

MMLU Pro

85.2%

91.1%

MMMLU

88.4%

93.0%

MRCR v2

66.4%

77.8%

SWE-bench Multilingual

51.7%

53.4%

SWE-bench Pro

35.7%

81.42%

SWE-bench Verified

52.0%

65.4%

Terminal-Bench 2.0

42.9%