わたしたちの信念

AA Coding Index

このベンチマークについて

Artificial Analysisが算出するコーディング能力総合指数。LiveCodeBench、SWE-bench、Aiderなど実際のコーディングベンチマークを総合します。

出典: Artificial Analysis
順位モデル
#1

OpenAI

GPT-5.4

57.3
#2

Google

Gemini 3.1 Pro

55.5
#3

Anthropic

Claude Opus 4.7

52.5
#4

Anthropic

Claude Opus 4.6

48.1
#5

Anthropic

Claude Opus 4.5

47.8
#6

DeepSeek

DeepSeek V4 Pro

47.5
#7

Meta

Muse Spark

47.5
#8

Z.ai

GLM-5

44.2
#9

Z.ai

GLM-5.1

43.4
#10

Anthropic

Claude Sonnet 4.6

43.0
#11

Alibaba

Qwen3.6 Plus

42.9
#12

Google

Gemini 3 Flash

42.6
#13

MiniMax

MiniMax M2.7

41.9
#14

Xiaomi

MiMo-V2-Pro

41.4
#15

Grok

Grok 4.20 (Reasoning)

40.5
#16

Moonshot AI

Kimi K2.5

39.5
#17

DeepSeek

DeepSeek V4 Flash

38.7
#18

Google

Gemma 4 31B

38.7
#19

Anthropic

Claude Sonnet 4.5

38.6
#20

OpenAI

GPT-5.4 Mini

37.5
#21

MiniMax

MiniMax M2.5

37.4
#22

Alibaba

Qwen3.5 397B A17B

37.4
#23

DeepSeek

DeepSeek V3.2

36.7
#24

Anthropic

Claude Opus 4.1

36.5
#25

OpenAI

GPT-5.4 Nano

35.0
#26

Anthropic

Claude Sonnet 4

34.1
#27

Anthropic

Claude Opus 4

34.0
#28

Anthropic

Claude Haiku 4.5

32.6
#29

Google

Gemini 2.5 Pro

31.9
#30

NVIDIA

Nemotron 3 Super

31.2
#31

Grok

Grok 4.1 Fast (Reasoning)

30.9
#32

Google

Gemini 3.1 Flash Lite

30.1
#33

Baidu

ERNIE 5.0 Thinking

29.2
#34

Arcee AI

Trinity Large Thinking

27.2
#35

Mistral AI

Mistral Small 4

24.3
#36

OpenAI

GPT-5 Nano

22.9
#37

Google

Gemini 2.5 Flash

22.2
#38

Grok

Grok 4.20

22.0
#39

OpenAI

GPT-5 Mini

21.9
#40

OpenAI

GPT-4.1

21.8
#41

OpenAI

GPT-5

21.2
#42

Grok

Grok 4.1 Fast

19.5
#43

Meituan

Longcat Flash Chat

16.5
#44

Meta

Llama 4 Maverick

15.6
#45

OpenAI

GPT OSS 120B

15.5
#46

Baidu

ERNIE 4.5 300B A47B

14.5
#47

Google

Gemini 2.5 Flash Lite

14.5
#48

Amazon

Nova 2 Lite

12.5
#49

Meta

Llama 4 Scout

6.7