大模型排行榜 · industry_legal_and_government

数据来自 LMArena（Chatbot Arena）：全球用户对匿名模型两两盲测投票，按 Elo 机制计算排名——相比跑分基准更接近真实使用体验。本页每周同步官方榜单，当前榜单发布于 2026-10-28。点击模型名可查看其各类别成绩与评分走势。

综合编程创意写作中文英文 exclude_ties expert french german hard_prompts hard_prompts_english industry_business_and_management_and_financial_o industry_entertainment_and_sports_and_media industry_legal_and_government industry_life_and_physical_and_social_science industry_mathematical industry_medicine_and_healthcare industry_software_and_it_services

排名	模型	开发方	Elo 评分	投票数
#1	Gemini 4 Carbon (High)	google	1550	—
#2	Gemini 4 Argon (High)	google	1524	—
#3	Muse Spark 1.2 (xHigh)	meta	1519	—
#4	Muse Wonder (Max)	meta	1513	—
#5	Claude Opus 4.7 (High)	anthropic	1511	—
#6	Claude Fable 5 (High)	anthropic	1510	—
#7	Kimi K3 (Max)	moonshot	1509	—
#8	Claude Opus 4.6 (High)	anthropic	1509	—
#9	Claude Opus 4.6	anthropic	1507	—
#10	GPT-6.1 Sol (Max)	openai	1505	—
#11	Claude Opus 5 (Max)	anthropic	1502	—
#12	Claude Opus 5 (High)	anthropic	1502	—
#13	Gemini 3 Pro	google	1501	—
#14	Muse Spark	meta	1500	—
#15	Claude Opus 5.5 (High)	anthropic	1500	—
#16	Muse Spark 1.3 (Max)	meta	1499	—
#17	Gemini 3.1 Pro Preview	google	1498	—
#18	Muse Spark 1.1	meta	1498	—
#19	Claude Opus 4.8	anthropic	1498	—
#20	Gemini 3.7 Flash (High)	google	1496	—
#21	Claude Opus 4.7	anthropic	1496	—
#22	GPT-5.5 (High)	openai	1496	—
#23	Claude Opus 4.8 (High)	anthropic	1495	—
#24	GPT-5.6 Sol (xHigh)	openai	1495	—
#25	GLM-5.2 (Max)	zai	1492	—
#26	Gemini 3.8 Flash (High)	google	1492	—
#27	Claude Fable 5.1 (Max)	anthropic	1492	—
#28	GPT-5.4 (High)	openai	1490	—
#29	Qwen3.8-Max	alibaba	1489	—
#30	Claude Opus 4.5 (High, 32K, 20251101)	anthropic	1487	—
#31	Claude Sonnet 5.5 (xHigh)	anthropic	1486	—
#32	GPT-5.5	openai	1486	—
#33	Claude Sonnet 4.6	anthropic	1484	—
#34	GPT-6 Astra (Max)	openai	1484	—
#35	Gemini 3.5 Flash (Medium)	google	1484	—
#36	Gemini 3 Flash	google	1483	—
#37	Claude Opus 4.5 (20251101)	anthropic	1483	—
#38	DeepSeek-V4-Pro (High, 20260813)	deepseek	1481	—
#39	GPT-5.2 Chat (20260210)	openai	1481	—
#40	Gemini 3.6 Flash (High)	google	1480	—
#41	Qwen3.7-Max-Preview	alibaba	1480	—
#42	Gemini 3.5 Flash (High)	google	1479	—
#43	Grok 4.20 Beta (Reasoning, 0309)	xai	1478	—
#44	Grok 4.5	xai	1477	—
#45	GPT-5.5 Instant	openai	1476	—
#46	GPT-5.4	openai	1475	—
#47	Grok 4.1 (Thinking)	xai	1475	—
#48	MiMo-V2.5-Pro	xiaomi	1475	—
#49	GPT-5.6 Terra (xHigh)	openai	1475	—
#50	DeepSeek-V4-Pro	deepseek	1474	—
#51	GLM-5.3 (Max)	zai	1474	—
#52	Qwen3.5-Max-Preview	alibaba	1474	—
#53	Grok 4.1	xai	1474	—
#54	Claude Sonnet 5 (High)	anthropic	1474	—
#55	DeepSeek-V4-Pro Preview (High)	deepseek	1473	—
#56	GLM-5.3-Flash	zai	1472	—
#57	GLM-5.1	zai	1471	—
#58	Grok 4.20 Multi-Agent Beta (0309)	xai	1470	—
#59	Gemini 2.5 Pro	google	1470	—
#60	MiMo-V2.6-Pro	xiaomi	1469	—
#61	GLM-5	zai	1469	—
#62	GPT-4.5 Preview (2025-02-27)	openai	1469	—
#63	Gemini 3 Flash (Minimal)	google	1468	—
#64	Grok 4.20 Beta 1	xai	1467	—
#65	GPT-5.3 Chat	openai	1467	—
#66	Dola Seed 2.0 Pro	bytedance	1466	—
#67	Claude Sonnet 4.5 (20250929)	anthropic	1465	—
#68	Kimi K2.6	moonshot	1464	—
#69	GPT-5.1 (High)	openai	1464	—
#70	Claude Sonnet 4.5 (High, 32K, 20250929)	anthropic	1464	—
#71	Grok 4.7 (xHigh)	xai	1464	—
#72	Claude Opus 4.1 (Thinking 16K, 20250805)	anthropic	1464	—
#73	Gemma 4 26B A4B	google	1464	—
#74	ERNIE 5.1	baidu	1463	—
#75	Qwen3.6-Max-Preview	alibaba	1463	—
#76	DeepSeek-V3.1-Terminus (Thinking)	deepseek	1461	—
#77	GPT-5.6 Luna (xHigh)	openai	1461	—
#78	ChatGPT-4o Latest (20250326)	openai	1460	—
#79	GPT-5 (High)	openai	1460	—
#80	DeepSeek-V4.1-Flash-20260910 (Max)	deepseek	1459	—
#81	GPT-5 Chat	openai	1459	—
#82	GPT-6 Sol (Max)	openai	1459	—
#83	Claude Opus 4.1 (20250805)	anthropic	1458	—
#84	DeepSeek-V4-Flash Preview (High)	deepseek	1458	—
#85	ERNIE-5.0-Preview (1203)	baidu	1457	—
#86	MiMo-V2-Pro	xiaomi	1457	—
#87	Gemini 3.5 Flash-Lite	google	1455	—
#88	GPT-5.4 mini (High)	openai	1455	—
#89	Qwen3.7-Plus	alibaba	1455	—
#90	Gemma 4 31B	google	1455	—
#91	GLM-4.7	zai	1454	—
#92	Grok 4.6 (High)	xai	1454	—
#93	o3 (2025-04-16)	openai	1452	—
#94	GPT-5.1	openai	1452	—
#95	Hy3	tencent	1450	—
#96	DeepSeek-V4-Flash	deepseek	1449	—
#97	Kimi K2 Preview (0905)	moonshot	1449	—
#98	GPT-6 Luna (Max)	openai	1448	—
#99	GPT-5.2 (High)	openai	1448	—
#100	ERNIE-5.0-Preview (1022)	baidu	1448	—
#101	amazon-nova-experimental-chat (26-02-10)	amazon	1447	—
#102	Qwen3.5-397B-A17B	alibaba	1446	—
#103	Kimi K2.5 (Thinking)	moonshot	1446	—
#104	Qwen3.6-Plus	alibaba	1446	—
#105	Step 5 Preview	stepfun	1446	—
#106	Hy3 preview	tencent	1446	—
#107	GPT-5.2	openai	1444	—
#108	Inkling	thinky	1441	—
#109	Gemini 3.1 Flash-Lite Preview	google	1440	—
#110	Mistral Medium 3.5	mistral	1440	—
#111	MiniMax-M3	minimax	1439	—
#112	Kimi K2.5 (Instant)	moonshot	1439	—
#113	MiMo-V2.6-Flash	xiaomi	1439	—
#114	ERNIE 5.0 (0110)	baidu	1439	—
#115	Grok 4.3	xai	1439	—
#116	Claude Opus 4 (20250514)	anthropic	1438	—
#117	GPT-4.1 (2025-04-14)	openai	1437	—
#118	MiMo-V2-Omni	xiaomi	1437	—
#119	Qwen3.8-27B	alibaba	1437	—
#120	Beam 1.0 501B A23B	reflection	1436	—
#121	GLM-4.6	zai	1436	—
#122	Kimi K2 Thinking Turbo	moonshot	1434	—
#123	Claude Opus 4 (Thinking 16K, 20250514)	anthropic	1433	—
#124	Qwen3-235B-A22B-Instruct (2507)	alibaba	1433	—
#125	GLM-5V-Turbo	zai	1433	—
#126	Muse Glimmer	meta	1433	—
#127	DeepSeek-V3.2	deepseek	1432	—
#128	Qwen3-Max-Preview	alibaba	1431	—
#129	Qwen3-VL-235B-A22B-Instruct	alibaba	1431	—
#130	MiMo-V2.5	xiaomi	1430	—
#131	DeepSeek-V3.2 (Thinking)	deepseek	1429	—
#132	LongCat-Flash-Chat Exp (2602)	meituan	1429	—
#133	Grok 4.1 Fast (Reasoning)	xai	1428	—
#134	Gemini 2.5 Flash	google	1426	—
#135	Gemini 2.5 Flash Preview (09-2025)	google	1426	—
#136	Grok 4 Fast (Reasoning)	xai	1426	—
#137	NVIDIA-Nemotron-3-Ultra-550B-A55B (NVFP4)	nvidia	1426	—
#138	Grok 3 Preview (02-24)	xai	1425	—
#139	Qwen3.5-122B-A10B	alibaba	1425	—
#140	DeepSeek-V3.2-Exp (Thinking)	deepseek	1424	—
#141	Mistral Large 4	mistral	1424	—
#142	Grok 4 (0709)	xai	1423	—
#143	Grok 4 Fast (Non-Reasoning)	xai	1422	—
#144	DeepSeek-V3.1 (Thinking)	deepseek	1422	—
#145	Mistral Large 3	mistral	1422	—
#146	DeepSeek-V3.1	deepseek	1419	—
#147	Kimi K2 Preview (0711)	moonshot	1418	—
#148	DeepSeek-V3.1-Terminus	deepseek	1418	—
#149	MiniMax-M2.7	minimax	1417	—
#150	GLM-4.5	zai	1417	—
#151	Mistral Medium 3.1 (2508)	mistral	1416	—
#152	Claude Haiku 4.5 (20251001)	anthropic	1416	—
#153	DeepSeek-R1 (0528)	deepseek	1415	—
#154	DeepSeek-V3.2-Exp	deepseek	1414	—
#155	Claude Haiku 5.5 (High)	anthropic	1413	—
#156	Qwen3-Max (2025-09-23)	alibaba	1411	—
#157	Mistral Medium 3 (2505)	mistral	1410	—
#158	DeepSeek-V3 (0324)	deepseek	1409	—
#159	Claude Sonnet 4 (20250514)	anthropic	1409	—
#160	Qwen3-Next-80B-A3B-Instruct	alibaba	1409	—
#161	o1 (2024-12-17)	openai	1407	—
#162	Qwen3.5-27B	alibaba	1406	—
#163	Claude Sonnet 4 (Thinking 32K, 20250514)	anthropic	1406	—
#164	Hunyuan T1 (20250711)	tencent	1406	—
#165	Qwen3-VL-235B-A22B-Thinking	alibaba	1405	—
#166	Qwen3-235B-A22B (Non-Thinking)	alibaba	1405	—
#167	MiniMax-M2.5	minimax	1404	—
#168	amazon-nova-experimental-chat (12-10)	amazon	1401	—
#169	GLM-4.6V	zai	1400	—
#170	GPT-5.4 nano (High)	openai	1400	—
#171	Qwen3.5-35B-A3B	alibaba	1400	—
#172	Qwen3-Coder-480B-A35B-Instruct	alibaba	1400	—
#173	Qwen3.5-Flash	alibaba	1399	—
#174	DeepSeek-R1	deepseek	1399	—
#175	o4-mini (2025-04-16)	openai	1399	—
#176	amazon-nova-experimental-chat (26-01-10)	amazon	1398	—
#177	Inkling-Small	thinky	1397	—
#178	Qwen3-235B-A22B-Thinking (2507)	alibaba	1397	—
#179	GPT-5 mini (High)	openai	1397	—
#180	Claude 3.7 Sonnet (Thinking 32K, 20250219)	anthropic	1397	—
#181	Gemini 2.5 Flash-Lite Preview (Non-Thinking, 09-2025)	google	1397	—
#182	GPT-4.1 mini (2025-04-14)	openai	1396	—
#183	MiniMax-M2.1 Preview	minimax	1396	—
#184	LongCat-Flash-Chat	meituan	1396	—
#185	MiMo-V2-Flash (Thinking)	xiaomi	1395	—
#186	Step-3.5-Flash	stepfun	1395	—
#187	Claude 3.7 Sonnet (20250219)	anthropic	1394	—
#188	Trinity Large Preview	—	1394	—
#189	MiMo-V2-Flash (Non-Thinking)	xiaomi	1391	—
#190	o1-preview	openai	1391	—
#191	NVIDIA-Nemotron-3.5-Super-VL-120B-A12B (NVFP4)	nvidia	1388	—
#192	Qwen2.5-Max	alibaba	1388	—
#193	Grok 3 Mini (High)	xai	1385	—
#194	Qwen3-30B-A3B-Instruct (2507)	alibaba	1384	—
#195	Gemma 3 27B IT	google	1384	—
#196	DeepSeek-V3	deepseek	1384	—
#197	Claude 3.5 Sonnet (20241022)	anthropic	1383	—
#198	Gemini 2.5 Flash-Lite Preview (Thinking, 06-17)	google	1383	—
#199	Gemma 3 12B IT	google	1383	—
#200	Hunyuan TurboS (20250416)	tencent	1383	—

说明：Elo 评分由 LMArena 根据真实用户盲测投票计算，分差越大代表偏好差距越显著；排名会随每期投票变化。本页仅同步展示并翻译整理，原始数据版权归 LMArena 所有。