생활·사회과학

이 벤치마크는?

과학 관련 전문가 프롬프트에서의 Arena Elo 점수입니다.

출처: Arena Intelligence
순위모델
#1

Anthropic

Claude Opus 4.6

1517
#2

Google

Gemini 3.1 Pro

1513
#3

Z.ai

GLM-5.1

1499
#4

Anthropic

Claude Sonnet 4.6

1492
#5

Anthropic

Claude Opus 4.5

1490
#6

Z.ai

GLM-5

1487
#7

Google

Gemini 3 Flash

1486
#8

Xiaomi

MiMo-V2-Pro

1486
#9

Grok

Grok 4.20

1485
#10

Grok

Grok 4.20 (Reasoning)

1485
#11

OpenAI

GPT-5.4

1480
#12

OpenAI

GPT-5.4 Pro

1480
#13

OpenAI

GPT-5.4 Mini

1476
#14

Google

Gemini 2.5 Pro

1474
#15

Moonshot AI

Kimi K2.5

1472
#16

Meituan

Longcat Flash Chat

1471
#17

Anthropic

Claude Opus 4.1

1471
#18

Anthropic

Claude Sonnet 4.5

1465
#19

Google

Gemma 4 31B

1464
#20

Google

Gemini 3.1 Flash Lite

1464
#21

Qwen

Qwen3.5 397B A17B

1459
#22

Baidu

ERNIE 5.0 Thinking

1458
#23

Grok

Grok 4.1 Fast

1450
#24

Grok

Grok 4.1 Fast (Reasoning)

1450
#25

DeepSeek

DeepSeek V3.2

1448
#26

OpenAI

GPT-5

1444
#27

Anthropic

Claude Opus 4

1440
#28

Google

Gemini 2.5 Flash

1429
#29

Anthropic

Claude Haiku 4.5

1421
#30

Anthropic

Claude Sonnet 4

1418
#31

OpenAI

GPT-5.4 Nano

1418
#32

MiniMax

MiniMax M2.5

1413
#33

MiniMax

MiniMax M2.7

1411
#34

Google

Gemini 2.5 Flash Lite

1403
#35

Arcee AI

Trinity Large Thinking

1402
#36

OpenAI

GPT-5 Mini

1396
#37

OpenAI

GPT OSS 120B

1362
#38

Amazon

Nova 2 Lite

1356
#39

OpenAI

GPT-5 Nano

1345
#40

Meta AI

Llama 4 Maverick

1344
#41

Meta AI

Llama 4 Scout

1336
#42

OpenAI

GPT-4.1

1322