Arena 종합

Name: Arena 종합
Creator: aib vote
License: https://www.gnu.org/licenses/gpl-3.0.html

이 벤치마크는?

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

순위	모델
#1	Anthropic Claude Opus 4.7	1503
#2	Anthropic Claude Opus 4.6	1503
#3	Google Gemini 3.1 Pro	1493
#4	Meta Muse Spark	1492
#5	Grok Grok 4.20	1482
#6	Grok Grok 4.20 (Reasoning)	1482
#7	OpenAI GPT-5.4	1481
#8	OpenAI GPT-5.4 Pro	1481
#9	Google Gemini 3 Flash	1474
#10	Anthropic Claude Opus 4.5	1473
#11	Z.ai GLM-5.1	1470
#12	DeepSeek DeepSeek V4 Pro	1463
#13	Anthropic Claude Sonnet 4.6	1463
#14	Moonshot AI Kimi K2.6	1458
#15	Z.ai GLM-5	1457
#16	OpenAI GPT-5.4 Mini	1457
#17	Anthropic Claude Sonnet 4.5	1453
#18	Google Gemma 4 31B	1451
#19	Baidu ERNIE 5.0 Thinking	1451
#20	Moonshot AI Kimi K2.5	1449
#21	Anthropic Claude Opus 4.1	1449
#22	Google Gemini 2.5 Pro	1448
#23	Xiaomi MiMo-V2-Pro	1448
#24	Alibaba Qwen3.6 Plus	1447
#25	Alibaba Qwen3.5 397B A17B	1446
#26	DeepSeek DeepSeek V4 Flash	1439
#27	Google Gemini 3.1 Flash Lite	1439
#28	Meituan Longcat Flash Chat	1434
#29	OpenAI GPT-5	1434
#30	Grok Grok 4.1 Fast	1432
#31	Grok Grok 4.1 Fast (Reasoning)	1432
#32	DeepSeek DeepSeek V3.2	1424
#33	Anthropic Claude Opus 4	1424
#34	Google Gemini 2.5 Flash	1411
#35	Anthropic Claude Haiku 4.5	1408
#36	OpenAI GPT-5.4 Nano	1405
#37	MiniMax MiniMax M2.7	1404
#38	MiniMax MiniMax M2.5	1400
#39	Anthropic Claude Sonnet 4	1399
#40	OpenAI GPT-5 Mini	1390
#41	Google Gemini 2.5 Flash Lite	1380
#42	Arcee AI Trinity Large Thinking	1375
#43	NVIDIA Nemotron 3 Super	1361
#44	OpenAI GPT OSS 120B	1353
#45	Amazon Nova 2 Lite	1337
#46	OpenAI GPT-5 Nano	1337
#47	Meta Llama 4 Maverick	1327
#48	Meta Llama 4 Scout	1322
#49	OpenAI GPT-4.1	1312