Arena 코딩 Elo

이 벤치마크는?

코딩 관련 프롬프트에서의 사용자 선호도를 집계한 Arena Elo 점수입니다.

출처: Arena Intelligence
순위모델
#1

Anthropic

Claude Opus 4.6

1555
#2

OpenAI

GPT-5.4

1532
#3

OpenAI

GPT-5.4 Pro

1532
#4

Google

Gemini 3.1 Pro

1531
#5

Anthropic

Claude Opus 4.5

1531
#6

Grok

Grok 4.20

1524
#7

Grok

Grok 4.20 (Reasoning)

1524
#8

Anthropic

Claude Sonnet 4.6

1522
#9

Z.ai

GLM-5.1

1520
#10

Anthropic

Claude Sonnet 4.5

1518
#11

Anthropic

Claude Opus 4.1

1513
#12

OpenAI

GPT-5.4 Mini

1512
#13

Meituan

Longcat Flash Chat

1511
#14

Moonshot AI

Kimi K2.5

1510
#15

Google

Gemini 3 Flash

1509
#16

Xiaomi

MiMo-V2-Pro

1505
#17

Anthropic

Claude Opus 4

1498
#18

Google

Gemma 4 31B

1497
#19

Baidu

ERNIE 5.0 Thinking

1494
#20

Z.ai

GLM-5

1489
#21

Qwen

Qwen3.5 397B A17B

1488
#22

Anthropic

Claude Haiku 4.5

1477
#23

Anthropic

Claude Sonnet 4

1473
#24

DeepSeek

DeepSeek V3.2

1468
#25

Google

Gemini 2.5 Pro

1467
#26

OpenAI

GPT-5

1467
#27

OpenAI

GPT-5.4 Nano

1465
#28

MiniMax

MiniMax M2.5

1464
#29

Google

Gemini 3.1 Flash Lite

1463
#30

Grok

Grok 4.1 Fast

1462
#31

Grok

Grok 4.1 Fast (Reasoning)

1462
#32

MiniMax

MiniMax M2.7

1460
#33

Arcee AI

Trinity Large Thinking

1443
#34

OpenAI

GPT-5 Mini

1429
#35

Google

Gemini 2.5 Flash

1425
#36

Google

Gemini 2.5 Flash Lite

1398
#37

Amazon

Nova 2 Lite

1397
#38

OpenAI

GPT OSS 120B

1391
#39

OpenAI

GPT-5 Nano

1382
#40

Meta AI

Llama 4 Maverick

1372
#41

Meta AI

Llama 4 Scout

1362
#42

OpenAI

GPT-4.1

1338