우리들의 비전

LiveBench 종합

이 벤치마크는?

LiveBench가 측정한 종합 점수 (0~100). 추론·코딩·수학·데이터 분석·언어·지시 따르기 6개 카테고리 평균입니다. 오염 없는 문제를 정기적으로 갱신해 최신 모델을 공정하게 평가합니다.

출처: LiveBench
순위모델
#1

OpenAI

GPT-5.5

81.3
#2

OpenAI

GPT-5.4

80.9
#3

Google

Gemini 3.1 Pro

80.7
#4

Anthropic

Claude Opus 4.6

76.8
#5

Moonshot AI

Kimi K2.6

72.4
#6

OpenAI

GPT-5

71.3
#7

Alibaba

Qwen3.6 Plus

70.8
#8

Z.ai

GLM-5.1

70.6
#9

Anthropic

Claude Sonnet 4.6

70.2
#10

Moonshot AI

Kimi K2.5

69.2
#11

Grok

Grok 4.20 (Reasoning)

69.0
#12

Z.ai

GLM-5

68.7
#13

Anthropic

Claude Opus 4.7

68.4
#14

MiniMax

MiniMax M2.7

65.0
#15

Google

Gemma 4 31B

62.4
#16

Google

Gemini 3.1 Flash Lite

62.1
#17

Anthropic

Claude Opus 4.1

61.4
#18

OpenAI

GPT-5 Mini

61.0
#19

Anthropic

Claude Sonnet 4

60.6
#20

MiniMax

MiniMax M2.5

60.3
#21

Grok

Grok 4.1 Fast (Reasoning)

60.1
#22

Xiaomi

MiMo-V2-Pro

58.4
#23

Google

Gemini 2.5 Pro

57.5
#24

Google

Gemini 3 Flash

54.4
#25

Anthropic

Claude Opus 4.5

53.6
#26

Anthropic

Claude Sonnet 4.5

51.3
#27

DeepSeek

DeepSeek V3.2

49.8
#28

OpenAI

GPT-5 Nano

48.0
#29

Google

Gemini 2.5 Flash

46.9
#30

OpenAI

GPT OSS 120B

46.4
#31

Anthropic

Claude Haiku 4.5

43.0
#32

Google

Gemini 2.5 Flash Lite

41.5
#33

Grok

Grok 4.20

37.9
#34

OpenAI

GPT-5.4 Mini

33.9
#35

NVIDIA

Nemotron 3 Super

32.0
#36

Grok

Grok 4.1 Fast

31.6
#37

Arcee AI

Trinity Large Thinking

30.4
#38

OpenAI

GPT-5.4 Nano

29.9