🏟️ Arena AI 排行榜

LMSYS 团队运营的真实人类盲投 ELO 榜 · 6 大维度 · 28,272,345.82 投票 · 自动 1 小时刷新

实时抓取缓存 1h

💬 Text Leaderboard · Top 30

通用文本对话 ELO 榜（盲投最多） · 414 模型（榜单共 414 名）·12,587,863 票

Rank	Model	Provider	Score (95% CI)	Votes	$ / M	Context
1	Google	gemini-4-argon-high	1525 ±9	4,892	$2 / $10	1M
2	claude-opus-5.5-high Anthropic	Anthropic	1507 ±8	6,272	$4 / $20	1M
3	claude-opus-4-6-high Anthropic	Anthropic	1504 ±3	79,415	$5 / $25	1M
4	claude-fable-5-high Anthropic	Anthropic	1504 ±4	40,026	$10 / $50	1M
5	claude-opus-4-7-high Anthropic	Anthropic	1501 ±4	65,192	$5 / $25	1M
6	claude-fable-5.1-max Anthropic	Anthropic	1501 ±6	13,063	$10 / $50	1M
7	claude-opus-4-6 Anthropic	Anthropic	1498 ±3	84,095	$5 / $25	1M
8	Google	gemini-3.8-flash-high	1497 ±5	31,437	$0.75 / $3.75	1M
9	muse-spark-1.3-max Meta	Meta	1494 ±6	14,694	$1.25 / $4.25	1M
10	claude-opus-4-7 Anthropic	Anthropic	1494 ±4	66,315	$5 / $25	1M
11	muse-spark-1.2 (xHigh) Meta	Meta	1492 ±8	6,118	$1.25 / $4.25	N/A
12	muse-spark-1.1 Meta	Meta	1491 ±4	39,847	$1.25 / $4.25	N/A
13	claude-opus-5-high Anthropic	Anthropic	1490 ±4	62,934	$5 / $25	1M
14	claude-opus-5-max Anthropic	Anthropic	1489 ±5	30,569	$5 / $25	1M
15	muse-spark Meta	Meta	1489 ±6	14,168	N/A	N/A
16	Moonshot	kimi-k3-max	1488 ±5	30,351	N/A	N/A
17	Google	gemini-3.1-pro-preview	1487 ±3	124,675	$1 / $6	1M
18	Google	gemini-3.7-flash-high	1486 ±5	24,173	$0.75 / $3.75	1M
19	Google	gemini-3-pro	1485 ±4	42,078	$2 / $12	1M
20	OpenAI	gpt-5.6-sol-xhigh	1485 ±4	38,490	$4 / $20	N/A
21	OpenAI	gpt-6.1-sol-max	1484 ±9	4,512	$2 / $10	1.1M
22	Alibaba	qwen3.8-max	1483 ±5	25,378	$2 / $6	1M
23	Google	gemini-3.6-flash-high	1482 ±4	36,172	$0.75 / $3.75	1M
24	OpenAI	gpt-5.5-high	1482 ±4	69,437	$5 / $30	1.1M
25	claude-opus-4-8-high Anthropic	Anthropic	1481 ±4	64,177	$5 / $25	1M
26	Xiaomi	mimo-v2.6-pro	1480 ±9	4,069	$0.43 / $0.87	1.1M
27	Z.ai	glm-5.3-max	1478 ±5	20,049	$1.40 / $4.40	1M
28	Google	gemini-3.5-flash-high	1478 ±4	48,533	$0.75 / $4.50	1M
29	OpenAI	gpt-5.5	1477 ±4	71,047	$5 / $30	1.1M
30	Google	gemini-3.5-flash-medium	1476 ±4	47,339	$0.75 / $4.50	1M

关于数据：Arena.ai 排行榜基于人类对模型两两盲投结果，由 LMSYS Chatbot Arena 团队运营，是 LLM 评测领域公信力最高的 ELO 排名之一。所有分数均含 95% 置信区间，投票数越大分数越稳定。本页 实时抓取 arena.ai/leaderboard/ 5 个端点（text / vision / agent / search / text-to-image，webdev 已下线），缓存 1 小时。