GPT-5.4 nano (xhigh) vs Claude 4.5 Haiku (Reasoning) | grube.ai

Back to Benchmarks

GPT-5.4 nano (xhigh)

OpenAI

vs

Claude 4.5 Haiku (Reasoning)

Anthropic

GPT-5.4 nano (xhigh) wins 5 of 11

Index Scores

44.4

+7.3

Intelligence

37.1

43.9

+11.3

Coding

32.6

0.0

Math

+83.7

83.7

Benchmarks

0.0

MMLU Pro

+76.0

76.0

81.7

+14.5

GPQA

67.2

26.5

+16.8

HLE

9.7

0.0

LiveCode

+61.5

61.5

46.9

+3.6

SciCode

43.3

0.0

MATH-500

0.0

0.0

AIME

0.0

Speed

65

TPS

+1

66

Shared

Winning delta

crafted by bart stefanski