AA Coding Index

이 벤치마크는?

Artificial Analysis가 산출하는 코딩 역량 종합 지수. LiveCodeBench, SWE-bench, Aider 등 실제 코드 생성 및 버그 수정 벤치마크를 종합합니다.

출처: Artificial Analysis
순위모델
#1

OpenAI

GPT-5.4

57.3
#2

Google

Gemini 3.1 Pro

55.5
#3

Anthropic

Claude Opus 4.6

48.1
#4

Anthropic

Claude Opus 4.5

47.8
#5

Meta AI

Muse Spark

47.5
#6

Z.ai

GLM-5

44.2
#7

Anthropic

Claude Sonnet 4.6

43.0
#8

Qwen

Qwen3.6 Plus

42.9
#9

Google

Gemini 3 Flash

42.6
#10

MiniMax

MiniMax M2.7

41.9
#11

Xiaomi

MiMo-V2-Pro

41.4
#12

Grok

Grok 4.20 (Reasoning)

40.5
#13

Moonshot AI

Kimi K2.5

39.5
#14

Google

Gemma 4 31B

38.7
#15

Anthropic

Claude Sonnet 4.5

38.6
#16

OpenAI

GPT-5.4 Mini

37.5
#17

MiniMax

MiniMax M2.5

37.4
#18

Qwen

Qwen3.5 397B A17B

37.4
#19

DeepSeek

DeepSeek V3.2

36.7
#20

Anthropic

Claude Opus 4.1

36.5
#21

OpenAI

GPT-5.4 Nano

35.0
#22

Anthropic

Claude Sonnet 4

34.1
#23

Anthropic

Claude Opus 4

34.0
#24

Anthropic

Claude Haiku 4.5

32.6
#25

Google

Gemini 2.5 Pro

31.9
#26

NVIDIA

Nemotron 3 Super

31.2
#27

Grok

Grok 4.1 Fast (Reasoning)

30.9
#28

Google

Gemini 3.1 Flash Lite

30.1
#29

Baidu

ERNIE 5.0 Thinking

29.2
#30

OpenAI

GPT-5 Nano

22.9
#31

Google

Gemini 2.5 Flash

22.2
#32

Grok

Grok 4.20

22.0
#33

OpenAI

GPT-5 Mini

21.9
#34

OpenAI

GPT-4.1

21.8
#35

OpenAI

GPT-5

21.2
#36

Grok

Grok 4.1 Fast

19.5
#37

Meituan

Longcat Flash Chat

16.5
#38

Meta AI

Llama 4 Maverick

15.6
#39

OpenAI

GPT OSS 120B

15.5
#40

Baidu

ERNIE 4.5 300B A47B

14.5
#41

Google

Gemini 2.5 Flash Lite

14.5
#42

Amazon

Nova 2 Lite

12.5
#43

Meta AI

Llama 4 Scout

6.7