中文大模型能力榜

📊 全部数据 · 总览

按 LMArena 中文竞技场(Chinese Arena)的真人中文盲测对战评分,排出大模型在中文场景的能力高下。它只统计中文对话里的偏好胜负,和上方「总榜」(以英文为主)可能明显不同——国产模型在这里通常更靠前。评分反映偏好、非绝对能力,头部分差常落在置信区间内(见下表),宜看作近似并列。数据归因 LMArena,不构成投资或采购建议。榜单发布于 2026-10-13

排名	模型	机构	Arena 分	置信区间	投票数
1	gemini-4-argon-high	google	1610	±32	361
2	claude-opus-5.5-high	anthropic	1585	±23	704
3	claude-fable-5.1-max	anthropic	1584	±19	1.1K
4	claude-opus-5-max	anthropic	1573	±13	2.5K
5	claude-opus-5-high	anthropic	1568	±9	5.2K
6	claude-opus-4-6-high	anthropic	1550	±9	5.6K
7	gemini-3.7-flash-high	google	1550	±14	2.0K
8	gemini-3.8-flash-high	google	1548	±12	2.9K
9	claude-opus-4-6	anthropic	1548	±9	5.8K
10	claude-fable-5-high	anthropic	1544	±11	3.1K
11	claude-sonnet-5.5-xhigh	anthropic	1538	±27	508
12	qwen3.8-max	alibaba	1534	±14	2.0K
13	qwen3.5-max-preview	alibaba	1534	±15	1.6K
14	gpt-5.5	openai	1533	±10	4.4K
15	claude-opus-4-7-high	anthropic	1532	±10	4.2K
16	claude-opus-4-7	anthropic	1531	±10	4.1K
17	gemini-3.6-flash-high	google	1530	±12	2.7K
18	glm-5.3-flash	zai	1530	±13	2.4K
19	kimi-k3-max	moonshot	1529	±13	2.3K
20	qwen3.7-max-preview	alibaba	1529	±34	316
21	mimo-v2.6-pro	xiaomi	1528	±33	299
22	glm-5.3-max	zai	1528	±15	1.6K
23	gemini-3.1-pro-preview	google	1528	±7	8.7K
24	muse-spark-1.3-max	meta	1528	±17	1.3K
25	gpt-5.6-sol-xhigh	openai	1526	±11	3.1K
26	gemini-3.5-flash-medium	google	1526	±11	3.4K
27	gpt-5.5-high	openai	1526	±10	4.5K
28	gemini-3.5-flash-high	google	1525	±11	3.6K
29	gemini-3-pro	google	1523	±10	3.5K
30	kimi-k2.6	moonshot	1520	±13	2.3K
31	glm-5.2-max	zai	1520	±11	3.5K
32	gpt-5.4-high	openai	1519	±10	4.1K
33	gpt-5.6-terra-xhigh	openai	1515	±11	3.3K
34	glm-5.1	zai	1514	±10	4.3K
35	muse-spark-1.1	meta	1514	±11	3.3K
36	ernie-5.0-preview-1022	baidu	1512	±30	390
37	glm-5	zai	1511	±13	2.0K
38	gemini-3-flash	google	1510	±12	2.5K
39	qwen3.7-plus	alibaba	1510	±12	2.6K
40	mimo-v2.6-flash	xiaomi	1510	±21	818
41	gpt-5.4	openai	1508	±10	4.4K
42	Step 5 Preview	stepfun	1508	±26	496
43	gemini-2.5-pro	google	1507	±7	9.3K
44	muse-spark	meta	1507	±20	974
45	ernie-5.1	baidu	1507	±13	2.1K
46	muse-spark-1.2 (xHigh)	meta	1507	±23	661
47	mimo-v2.5-pro	xiaomi	1507	±9	4.8K
48	claude-opus-4-8-high	anthropic	1506	±10	4.4K
49	claude-opus-4-8	anthropic	1506	±10	4.4K
50	qwen3.8-27b	alibaba	1503	±15	1.7K
51	deepseek-v4.1-flash-max	deepseek	1501	±20	918
52	ernie-5.0-preview-1203	baidu	1501	±21	830
53	glm-4.6	zai	1499	±11	2.7K
54	qwen3.5-397b-a17b	alibaba	1499	±8	6.4K
55	nvidia-nemotron-3-ultra-550b-a55b-nvfp4	nvidia	1497	±22	725
56	glm-4.7	zai	1495	±18	1.0K
57	gpt-5.1-high	openai	1495	±10	3.4K
58	mistral-large-4	mistral	1495	±24	589
59	gemma-4-26b-a4b	google	1494	±25	524
60	kimi-k2.5-thinking	moonshot	1494	±9	4.6K
61	grok-4.5	xai	1494	±11	3.1K
62	gpt-6.1-sol-max	openai	1494	±25	536
63	ernie-5.0-0110	baidu	1492	±12	2.5K
64	gpt-6-astra-max	openai	1492	±20	877
65	claude-sonnet-4-6	anthropic	1491	±9	4.5K
66	inkling	thinky	1491	±11	3.1K
67	dola-seed-2.0-pro	bytedance	1489	±9	5.3K
68	glm-5v-turbo	zai	1489	±21	764
69	qwen3.6-max-preview	alibaba	1488	±31	358
70	hy3	tencent	1487	±20	894
71	deepseek-v4-pro-high-preview	deepseek	1486	±11	3.3K
72	deepseek-v4-pro	deepseek	1481	±11	3.4K
73	qwen3-235b-a22b-thinking-2507	alibaba	1481	±26	536
74	grok-4.20-beta-0309-reasoning	xai	1481	±10	4.2K
75	deepseek-v4-pro-high-20260813	deepseek	1480	±22	704
76	claude-sonnet-5-high	anthropic	1478	±10	3.7K
77	qwen3-max-preview	alibaba	1478	±13	2.1K
78	qwen3.5-27b	alibaba	1478	±13	2.1K
79	gemini-3-flash (thinking-minimal)	google	1477	±8	6.0K
80	grok-4.20-beta1	xai	1476	±15	1.7K
81	grok-4.6-high	xai	1476	±13	2.2K
82	qwen3.6-plus	alibaba	1475	±12	2.7K
83	gemma-4-31b	google	1475	±25	522
84	grok-4.20-multi-agent-beta-0309	xai	1475	±10	4.0K
85	gpt-5.6-luna-xhigh	openai	1470	±11	3.2K
86	claude-opus-4-5-20251101	anthropic	1470	±8	5.4K
87	deepseek-v3.1-thinking	deepseek	1469	±18	1.1K
88	deepseek-v4-flash-high-preview	deepseek	1468	±11	3.0K
89	mimo-v2.5	xiaomi	1468	±12	2.6K
90	gemini-3.5-flash-lite	google	1467	±12	2.7K
91	minimax-m3	minimax	1467	±10	4.0K
92	mimo-v2-omni	xiaomi	1467	±19	1.0K
93	claude-opus-4-5-20251101-high-32k	anthropic	1467	±11	3.0K
94	gpt-5.1	openai	1466	±10	3.6K
95	glm-4.5	zai	1465	±14	1.8K
96	qwen3-235b-a22b-instruct-2507	alibaba	1465	±7	7.3K
97	amazon-nova-experimental-chat-26-02-10	amazon	1464	±32	316
98	longcat-flash-chat-2602-exp	meituan	1464	±14	2.0K
99	grok-4.1	xai	1464	±9	5.3K
100	deepseek-v4-flash	deepseek	1464	±11	3.0K

能力评分数据来自 LMArena(原 LMSYS Chatbot Arena)公开发布的排行榜数据集「中文(chinese)」切片,许可 CC-BY-4.0,本站署名引用、未作修改。 Arena 分由真人匿名中文盲测对战经统计模型(Bradley-Terry)估算,反映相对偏好、非绝对能力,可能滞后或随样本波动,仅供参考。本栏目仅作信息整理,不构成任何投资或采购建议。本站不自建评测跑分。榜单发布于 2026-10-13。