Arena 종합

이 벤치마크는?

수백만 건의 실제 사용자 선호도 투표를 Bradley-Terry 모델로 집계한 종합 Elo 점수. 점수가 높을수록 실제 사용자가 더 선호하는 모델입니다.

출처: Arena Intelligence
순위모델
#1

Anthropic

Claude Opus 4.6

1503
#2

Google

Gemini 3.1 Pro

1493
#3

Grok

Grok 4.20

1490
#4

Grok

Grok 4.20 (Reasoning)

1490
#5

ChatGPT

GPT-5.4

1484
#6

ChatGPT

GPT-5.4 Pro

1484
#7

Google

Gemini 3 Flash

1474
#8

Anthropic

Claude Opus 4.5

1473
#9

Z.ai

GLM-5.1

1467
#10

Anthropic

Claude Sonnet 4.6

1462
#11

ChatGPT

GPT-5.4 Mini

1458
#12

Z.ai

GLM-5

1456
#13

Moonshot AI

Kimi K2.5

1453
#14

Anthropic

Claude Sonnet 4.5

1452
#15

Baidu

ERNIE 5.0 Thinking

1451
#16

Google

Gemma 4 31B

1451
#17

Anthropic

Claude Opus 4.1

1449
#18

Google

Gemini 2.5 Pro

1448
#19

Qwen

Qwen3.5 397B A17B

1448
#20

Xiaomi

MiMo-V2-Pro

1445
#21

Meituan

Longcat Flash Chat

1444
#22

Google

Gemini 3.1 Flash Lite

1436
#23

ChatGPT

GPT-5

1433
#24

Grok

Grok 4.1 Fast

1431
#25

Grok

Grok 4.1 Fast (Reasoning)

1431
#26

DeepSeek

DeepSeek V3.2

1424
#27

Anthropic

Claude Opus 4

1424
#28

Google

Gemini 2.5 Flash

1411
#29

Anthropic

Claude Haiku 4.5

1407
#30

MiniMax

MiniMax M2.5

1404
#31

MiniMax

MiniMax M2.7

1403
#32

ChatGPT

GPT-5.4 Nano

1403
#33

Anthropic

Claude Sonnet 4

1399
#34

ChatGPT

GPT-5 Mini

1390
#35

Google

Gemini 2.5 Flash Lite

1380
#36

Arcee AI

Trinity Large Thinking

1375
#37

ChatGPT

GPT OSS 120B

1354
#38

Amazon

Nova 2 Lite

1338
#39

ChatGPT

GPT-5 Nano

1336
#40

Meta AI

Llama 4 Maverick

1327
#41

Meta AI

Llama 4 Scout

1322
#42

ChatGPT

GPT-4.1

1312