GPT-5.4 (xhigh) vs Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort) | grube.ai

Back to Benchmarks

GPT-5.4 (xhigh)

OpenAI

vs

Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)

Anthropic

GPT-5.4 (xhigh) wins 5 of 11

Index Scores

57.2

+5.5

Intelligence

51.7

57.3

+6.4

Coding

50.9

0.0

Math

0.0

Benchmarks

0.0

MMLU Pro

0.0

92.0

+4.5

GPQA

87.5

41.6

+11.6

HLE

30.0

0.0

LiveCode

0.0

56.6

+9.8

SciCode

46.8

0.0

MATH-500

0.0

0.0

AIME

0.0

Speed

32

TPS

+10

42

Shared

Winning delta

crafted by bart stefanski