수학·통계

이 벤치마크는?

수학 관련 전문가 프롬프트에서의 Arena Elo 점수입니다.

출처: Arena Intelligence
순위모델
#1

Anthropic

Claude Opus 4.6

1516
#2

OpenAI

GPT-5.4

1511
#3

OpenAI

GPT-5.4 Pro

1511
#4

Google

Gemini 3.1 Pro

1494
#5

Anthropic

Claude Sonnet 4.6

1487
#6

Moonshot AI

Kimi K2.5

1485
#7

Google

Gemini 3 Flash

1472
#8

Anthropic

Claude Opus 4.5

1470
#9

Grok

Grok 4.20

1470
#10

Grok

Grok 4.20 (Reasoning)

1470
#11

Google

Gemma 4 31B

1470
#12

Z.ai

GLM-5

1469
#13

Anthropic

Claude Sonnet 4.5

1468
#14

Z.ai

GLM-5.1

1466
#15

OpenAI

GPT-5.4 Mini

1459
#16

Xiaomi

MiMo-V2-Pro

1455
#17

Google

Gemini 2.5 Pro

1450
#18

Anthropic

Claude Opus 4.1

1450
#19

Qwen

Qwen3.5 397B A17B

1450
#20

OpenAI

GPT-5

1442
#21

OpenAI

GPT-5.4 Nano

1441
#22

Google

Gemini 3.1 Flash Lite

1441
#23

Meituan

Longcat Flash Chat

1440
#24

DeepSeek

DeepSeek V3.2

1439
#25

Baidu

ERNIE 5.0 Thinking

1437
#26

MiniMax

MiniMax M2.5

1430
#27

Anthropic

Claude Opus 4

1423
#28

Grok

Grok 4.1 Fast

1422
#29

Grok

Grok 4.1 Fast (Reasoning)

1422
#30

Google

Gemini 2.5 Flash

1420
#31

MiniMax

MiniMax M2.7

1419
#32

Anthropic

Claude Haiku 4.5

1416
#33

Anthropic

Claude Sonnet 4

1409
#34

OpenAI

GPT-5 Mini

1409
#35

OpenAI

GPT OSS 120B

1385
#36

Arcee AI

Trinity Large Thinking

1380
#37

Google

Gemini 2.5 Flash Lite

1370
#38

OpenAI

GPT-5 Nano

1351
#39

Amazon

Nova 2 Lite

1349
#40

Meta AI

Llama 4 Maverick

1320
#41

Meta AI

Llama 4 Scout

1313
#42

OpenAI

GPT-4.1

1308