Text Generation Arena 文本生成模型排行榜
基于 Text Generation Arena 用户匿名投票的最新AI文本生成模型排行榜，涵盖各模型的 Elo 得分、95% 置信区间、投票量、机构与许可证。
榜首模型
gemini-4-argon-high
最高得分
1,525
模型数量
410
数据版本
2026年09月30日
数据来源: LM Arena
关于本排行榜
本排行榜展示了当前最强 AI 大模型在文本生成任务中的综合实力排名。数据来源于 LMArena（前身为 LMSYS Chatbot Arena），这是目前全球最大的 AI 模型众包评测平台。用户在平台上与两个匿名模型同时对话，并投票选出更好的回答——排名完全由真实用户的偏好决定，而非实验室基准测试。
评测方法概要
匿名盲测：用户同时与两个"隐藏身份"的模型对话，根据回答质量投票，排除品牌偏见。
Elo 评分：基于国际象棋领域的 Elo Rating 体系（Bradley-Terry 模型），通过对战结果计算每个模型的实力分数。分数越高，说明模型在真实对话中被用户选中的概率越大。
场景覆盖广泛：涵盖编程、创意写作、数学推理、知识问答、角色扮演等高频真实场景。
DataLearner 在原始数据基础上提供中文解读与深度分析，并将排行榜模型关联至 DataLearner 模型库，方便您一键查看模型详情、API 定价、评测得分等完整信息。
排名总表
| 排名 | 模型名称 | 得分 | 95% CI | 投票数 | 机构 | 许可证 |
|---|---|---|---|---|---|---|
 gemini-4-argon-highGoogle | 1,525 | +/-9 | 4,942 | Proprietary | ||
 Claude Opus 4.6 (high)Anthropic | 1,505 | +/-3 | 77,193 | Anthropic | Proprietary | |
| claude-fable-5-highAnthropic | 1,505 | +/-4 | 37,900 | Anthropic | Proprietary | |
| 4 | claude-opus-5.5-highAnthropic | 1,504 | +/-10 | 3,932 | Anthropic | Proprietary |
| 5 | Opus 4.7 (high)Anthropic | 1,502 | +/-4 | 64,607 | Anthropic | Proprietary |
| 6 | Claude Fable 5.1 (max)Anthropic | 1,501 | +/-7 | 11,241 | Anthropic | Proprietary |
| 7 | Claude Opus 4.6Anthropic | 1,497 | +/-3 | 81,769 | Anthropic | Proprietary |
| 8 |  muse-spark-1.3-maxMeta | 1,495 | +/-6 | 11,698 | Meta | Proprietary |
| 9 | Opus 4.7Anthropic | 1,494 | +/-4 | 65,705 | Anthropic | Proprietary |
| 10 | Muse Spark 1.2 (xhigh)Facebook AI研究实验室 | 1,494 | +/-10 | 3,833 | Facebook AI研究实验室 | Proprietary |
| 11 | gemini-3.8-flash-highGoogle | 1,494 | +/-5 | 24,828 | Proprietary | |
| 12 | Muse Spark 1.1Facebook AI研究实验室 | 1,492 | +/-4 | 36,558 | Facebook AI研究实验室 | Proprietary |
| 13 | Claude Opus 5 (high)Anthropic | 1,491 | +/-4 | 58,380 | Anthropic | Proprietary |
| 14 | Claude Opus 5 (max)Anthropic | 1,489 | +/-5 | 28,351 | Anthropic | Proprietary |
| 15 | Muse SparkFacebook AI研究实验室 | 1,489 | +/-6 | 14,128 | Facebook AI研究实验室 | Proprietary |
| 16 | gemini-3.7-flash-highGoogle | 1,488 | +/-5 | 20,851 | Proprietary | |
| 17 |  Kimi K3 (max)Moonshot AI | 1,488 | +/-5 | 27,719 | Moonshot AI | Kimi K3 license |
| 18 | Gemini 3.1 Pro PreviewGoogle DeepMind | 1,487 | +/-3 | 121,225 | Google DeepMind | Proprietary |
| 19 | Gemini 3 ProGoogle DeepMind | 1,485 | +/-4 | 41,910 | Google DeepMind | Proprietary |
| 20 |  GPT-5.6 Sol (xhigh)OpenAI | 1,484 | +/-4 | 36,002 | OpenAI | Proprietary |
| 21 | gemini-3.6-flash-highGoogle | 1,483 | +/-4 | 35,640 | Proprietary | |
| 22 | GPT-5.5 (high)OpenAI | 1,481 | +/-4 | 69,312 | OpenAI | Proprietary |
| 23 | Claude Opus 4.8 (high)Anthropic | 1,481 | +/-4 | 63,577 | Anthropic | Proprietary |
| 24 |  Qwen3.8-Max阿里巴巴 | 1,481 | +/-5 | 22,809 | 阿里巴巴 | Proprietary |
| 25 | mimo-v2.6-proXiaomi | 1,480 | +/-9 | 4,074 | Xiaomi | MIT |
| 26 |  glm-5.3-maxZ.ai | 1,479 | +/-6 | 17,268 | Z.ai | MIT |
| 27 | gemini-3.5-flash-highGoogle | 1,477 | +/-4 | 47,981 | Proprietary | |
| 28 | GPT-5.5OpenAI | 1,477 | +/-4 | 70,894 | OpenAI | Proprietary |
| 29 | gemini-3.5-flash-mediumGoogle | 1,476 | +/-4 | 46,735 | Proprietary | |
| 30 | gpt-6-astra-maxOpenAI | 1,476 | +/-7 | 8,565 | OpenAI | Proprietary |
| 31 | GPT-5.2 Chat (0210)OpenAI | 1,476 | +/-4 | 35,943 | OpenAI | Proprietary |
| 32 | glm-5.2-maxZ.ai | 1,475 | +/-4 | 44,949 | Z.ai | MIT |
| 33 | GPT-5.4 (high)OpenAI | 1,475 | +/-4 | 64,648 | OpenAI | Proprietary |
| 34 |  Grok 4.20 BetaxAI | 1,475 | +/-5 | 27,840 | xAI | Proprietary |
| 35 | Qwen3.7-Max-Preview阿里巴巴 | 1,475 | +/-10 | 3,924 | 阿里巴巴 | Proprietary |
| 36 | Claude Opus 4.8Anthropic | 1,475 | +/-4 | 64,711 | Anthropic | Proprietary |
| 37 | Opus 4.5Anthropic | 1,474 | +/-4 | 37,446 | Anthropic | Proprietary |
| 38 | glm-5.3-flashZ.ai | 1,474 | +/-5 | 21,644 | Z.ai | MIT |
| 39 | GPT-5.5 InstantOpenAI | 1,473 | +/-5 | 27,101 | OpenAI | Proprietary |
| 40 |  deepseek-v4.1-flash-maxDeepSeek | 1,473 | +/-7 | 8,125 | DeepSeek | MIT |
| 41 | Gemini 3.0 FlashGoogle DeepMind | 1,473 | +/-4 | 31,255 | Google DeepMind | Proprietary |
| 42 | Claude Sonnet 4.6Anthropic | 1,472 | +/-4 | 70,828 | Anthropic | Proprietary |
| 43 | grok-4.20-beta-0309-reasoningxAI | 1,472 | +/-4 | 66,372 | xAI | Proprietary |
| 44 | grok-4.20-multi-agent-beta-0309xAI | 1,471 | +/-4 | 64,838 | xAI | Proprietary |
| 45 | Claude Opus 4Anthropic | 1,470 | +/-3 | 72,799 | Anthropic | Proprietary |
| 46 |  ERNIE-5.1-Preview百度 | 1,468 | +/-5 | 39,421 | 百度 | Proprietary |
| 47 | mimo-v2.5-proXiaomi | 1,467 | +/-4 | 70,317 | Xiaomi | MIT |
| 48 | Grok 4.5xAI | 1,466 | +/-4 | 39,339 | xAI | Proprietary |
| 49 | Grok 4.1 ThinkingxAI | 1,465 | +/-3 | 66,373 | xAI | Proprietary |
| 50 | GPT-5.6 Terra (xhigh)OpenAI | 1,465 | +/-4 | 37,056 | OpenAI | Proprietary |
| 51 | GPT-5.4OpenAI | 1,465 | +/-4 | 67,753 | OpenAI | Proprietary |
| 52 | Qwen3.5 Max Preview阿里巴巴 | 1,465 | +/-5 | 22,566 | 阿里巴巴 | Proprietary |
| 53 | GLM 5.1智谱AI | 1,465 | +/-4 | 57,877 | 智谱AI | MIT |
| 54 | deepseek-v4-pro-high-20260813DeepSeek | 1,464 | +/-7 | 9,825 | DeepSeek | MIT |
| 55 | Claude Sonnet 4.5 (high)Anthropic | 1,462 | +/-4 | 45,191 | Anthropic | Proprietary |
| 56 | Kimi K2.6Moonshot AI | 1,461 | +/-4 | 39,950 | Moonshot AI | Modified MIT |
| 57 | Qwen3.6-Max-Preview阿里巴巴 | 1,460 | +/-8 | 5,408 | 阿里巴巴 | Proprietary |
| 58 | Grok 4.1xAI | 1,459 | +/-3 | 68,498 | xAI | Proprietary |
| 59 | Gemini 3.0 Flash (minimal)Google DeepMind | 1,458 | +/-3 | 90,888 | Google DeepMind | Proprietary |
| 60 | DeepSeek-V4-ProDeepSeek-AI | 1,458 | +/-4 | 57,549 | DeepSeek-AI | MIT |
| 61 | GLM-5智谱AI | 1,458 | +/-4 | 29,203 | 智谱AI | MIT |
| 62 |  Hy3腾讯AI实验室 | 1,457 | +/-7 | 10,239 | 腾讯AI实验室 | Apache 2.0 |
| 63 | Claude Sonnet 4.5Anthropic | 1,457 | +/-3 | 83,682 | Anthropic | Proprietary |
| 64 |  DOLA Seed 2.0 Pro字节跳动Seed团队 | 1,457 | +/-3 | 78,956 | 字节跳动Seed团队 | Proprietary |
| 65 | Qwen3.7-Plus阿里巴巴 | 1,456 | +/-4 | 42,001 | 阿里巴巴 | Proprietary |
| 66 | gpt-6-sol-maxOpenAI | 1,456 | +/-7 | 7,574 | OpenAI | Proprietary |
| 67 | GPT-5.1 Pro (high)OpenAI | 1,456 | +/-4 | 41,568 | OpenAI | Proprietary |
| 68 | Claude Sonnet 4.5Anthropic | 1,455 | +/-3 | 82,438 | Anthropic | Proprietary |
| 69 | deepseek-v4-pro-high-previewDeepSeek | 1,455 | +/-4 | 54,823 | DeepSeek | MIT |
| 70 | Gemini 3.5 Flash-LiteGoogle DeepMind | 1,454 | +/-4 | 35,335 | Google DeepMind | Proprietary |
| 71 | mimo-v2.6-flashXiaomi | 1,454 | +/-8 | 5,465 | Xiaomi | MIT |
| 72 | GPT-5.6 Luna (xhigh)OpenAI | 1,453 | +/-4 | 37,593 | OpenAI | Proprietary |
| 73 | grok-4.6-highSpaceXAI | 1,453 | +/-5 | 23,012 | SpaceXAI | Proprietary |
| 74 | Gemma 4 31BGoogle DeepMind | 1,453 | +/-7 | 6,137 | Google DeepMind | Apache 2.0 |
| 75 | Kimi K2 ThinkingMoonshot AI | 1,450 | +/-3 | 74,623 | Moonshot AI | Modified MIT |
| 76 | GPT-5.3 ChatOpenAI | 1,450 | +/-4 | 34,303 | OpenAI | Proprietary |
| 77 | Opus 4.1 (thinking-16k)Anthropic | 1,450 | +/-3 | 49,334 | Anthropic | Proprietary |
| 78 | ERNIE 5.0 Preview (1203)百度 | 1,449 | +/-7 | 9,788 | 百度 | Proprietary |
| 79 | Opus 4.1Anthropic | 1,448 | +/-3 | 76,927 | Anthropic | Proprietary |
| 80 | mimo-v2-proXiaomi | 1,448 | +/-5 | 25,397 | Xiaomi | Proprietary |
| 81 | GPT-5.4 mini (high)OpenAI | 1,447 | +/-4 | 63,311 | OpenAI | Proprietary |
| 82 | ERNIE 5.0 (0110)百度 | 1,447 | +/-4 | 36,422 | 百度 | Proprietary |
| 83 | Gemini 2.5 ProGoogle DeepMind | 1,446 | +/-2 | 124,859 | Google DeepMind | Proprietary |
| 84 | GPT-4.5OpenAI | 1,445 | +/-6 | 14,547 | OpenAI | Proprietary |
| 85 | gpt-6-luna-maxOpenAI | 1,444 | +/-7 | 7,654 | OpenAI | Proprietary |
| 86 | Qwen 3.6 Plus Preview阿里巴巴 | 1,443 | +/-4 | 48,092 | 阿里巴巴 | Proprietary |
| 87 | GPT-4o(2025-03-27)OpenAI | 1,443 | +/-3 | 81,601 | OpenAI | Proprietary |
| 88 | grok-4.7-xhighSpaceXAI | 1,442 | +/-8 | 5,675 | SpaceXAI | Proprietary |
| 89 | InklingThinking Machines Lab | 1,442 | +/-5 | 34,594 | Thinking Machines Lab | Apache 2.0 |
| 90 | Qwen3.5-397B-A17B阿里巴巴 | 1,442 | +/-3 | 88,092 | 阿里巴巴 | Apache 2.0 |
| 91 | GLM-4.7智谱AI | 1,442 | +/-6 | 12,219 | 智谱AI | MIT |
| 92 | Grok 4.3 BetaxAI | 1,441 | +/-4 | 71,269 | xAI | Proprietary |
| 93 |  MiniMax M3MiniMaxAI | 1,440 | +/-4 | 58,171 | MiniMaxAI | MiniMax Community License |
| 94 | qwen3.8-27bAlibaba | 1,439 | +/-5 | 17,777 | Alibaba | Apache 2.0 |
| 95 | GPT-5.1OpenAI | 1,438 | +/-4 | 44,311 | OpenAI | Proprietary |
| 96 | deepseek-v4-flash-high-previewDeepSeek | 1,438 | +/-4 | 51,691 | DeepSeek | MIT |
| 97 | Gemma 4 26B A4BGoogle DeepMind | 1,437 | +/-7 | 6,083 | Google DeepMind | Apache 2.0 |
| 98 | GPT-5.2 Pro (high)OpenAI | 1,437 | +/-4 | 49,549 | OpenAI | Proprietary |
| 99 |  longcat-flash-chat-2602-expMeituan | 1,437 | +/-5 | 29,235 | Meituan | Proprietary |
| 100 | DeepSeek-V4-FlashDeepSeek-AI | 1,436 | +/-4 | 51,869 | DeepSeek-AI | MIT |
| 101 | GPT-5.2OpenAI | 1,436 | +/-3 | 83,682 | OpenAI | Proprietary |
| 102 | GPT-5-Pro (high)OpenAI | 1,434 | +/-5 | 31,487 | OpenAI | Proprietary |
| 103 | Qwen3 Max (Preview)阿里巴巴 | 1,434 | +/-5 | 27,239 | 阿里巴巴 | Proprietary |
| 104 | GLM-5V-Turbo智谱AI | 1,434 | +/-6 | 10,448 | 智谱AI | Proprietary |
| 105 | mimo-v2.5Xiaomi | 1,434 | +/-4 | 47,292 | Xiaomi | MIT |
| 106 | Gemini 3.1 Flash-LiteGoogle DeepMind | 1,433 | +/-4 | 64,094 | Google DeepMind | Proprietary |
| 107 | OpenAI o3OpenAI | 1,432 | +/-4 | 58,598 | OpenAI | Proprietary |
| 108 | mimo-v2-omniXiaomi | 1,431 | +/-6 | 20,611 | Xiaomi | Proprietary |
| 109 | Kimi K2.5 InstantMoonshot AI | 1,431 | +/-7 | 8,355 | Moonshot AI | Modified MIT |
| 110 | Grok 4.1 Fast (fast-reasoning)xAI | 1,430 | +/-3 | 57,425 | xAI | Proprietary |
| 111 | Kimi K2 Thinking (thinking-turbo)Moonshot AI | 1,430 | +/-3 | 63,225 | Moonshot AI | Modified MIT |
| 112 |  Mistral Medium 3.5MistralAI | 1,427 | +/-6 | 11,757 | MistralAI | Modified MIT |
| 113 | GPT-5.2 Chat (0210)OpenAI | 1,427 | +/-4 | 31,017 | OpenAI | Proprietary |
| 114 |  amazon-nova-experimental-chat-26-02-10Amazon | 1,427 | +/-10 | 3,519 | Amazon | Proprietary |
| 115 |  Nemotron 3 UltraNVIDIA | 1,426 | +/-7 | 11,532 | NVIDIA | OpenMDW-1.1 |
| 116 | Claude Opus 4 (thinking-16k)Anthropic | 1,426 | +/-4 | 35,693 | Anthropic | Proprietary |
| 117 | Muse Glimmer-30BFacebook AI研究实验室 | 1,425 | +/-10 | 3,897 | Facebook AI研究实验室 | Apache-2.0 |
| 118 | DeepSeek V3.2-Exp (thinking)DeepSeek-AI | 1,425 | +/-7 | 8,948 | DeepSeek-AI | MIT |
| 119 | DeepSeek V3.2DeepSeek-AI | 1,425 | +/-3 | 48,004 | DeepSeek-AI | MIT |
| 120 | GLM-4.6智谱AI | 1,424 | +/-4 | 35,668 | 智谱AI | MIT |
| 121 | qwen3-max-2025-09-23Alibaba | 1,423 | +/-7 | 8,969 | Alibaba | Proprietary |
| 122 | DeepSeek V3.2-Exp (thinking)DeepSeek-AI | 1,423 | +/-4 | 41,966 | DeepSeek-AI | MIT |
| 123 | Qwen3-235B-A22B-2507阿里巴巴 | 1,422 | +/-3 | 97,808 | 阿里巴巴 | Apache 2.0 |
| 124 | DeepSeek V3.2-ExpDeepSeek-AI | 1,422 | +/-6 | 11,943 | DeepSeek-AI | MIT |
| 125 | DeepSeek-R1-0528DeepSeek-AI | 1,422 | +/-6 | 18,060 | DeepSeek-AI | MIT |
| 126 | Grok 4 FastxAI | 1,419 | +/-8 | 6,323 | xAI | Proprietary |
| 127 | ERNIE 5.0百度 | 1,419 | +/-9 | 4,724 | 百度 | Proprietary |
| 128 | Kimi K2 0905Moonshot AI | 1,419 | +/-7 | 11,539 | Moonshot AI | Modified MIT |
| 129 | Kimi K2Moonshot AI | 1,418 | +/-5 | 26,967 | Moonshot AI | Modified MIT |
| 130 | DeepSeek-V3.1DeepSeek-AI | 1,417 | +/-6 | 14,540 | DeepSeek-AI | MIT |
| 131 | DeepSeek-V3.1 Terminus (thinking)DeepSeek-AI | 1,417 | +/-10 | 3,357 | DeepSeek-AI | MIT |
| 132 | Qwen3.5-122B-A10B阿里巴巴 | 1,416 | +/-4 | 29,666 | 阿里巴巴 | Apache 2.0 |
| 133 | DeepSeek-V3.1 (thinking)DeepSeek-AI | 1,416 | +/-7 | 11,413 | DeepSeek-AI | MIT |
| 134 | MiniMax-M2.7MiniMaxAI | 1,415 | +/-3 | 79,403 | MiniMaxAI | Modified MIT |
| 135 | GPT-4.1OpenAI | 1,415 | +/-4 | 50,007 | OpenAI | Proprietary |
| 136 | DeepSeek-V3.1 TerminusDeepSeek-AI | 1,415 | +/-10 | 3,602 | DeepSeek-AI | MIT |
| 137 | Claude Opus 4Anthropic | 1,414 | +/-4 | 42,804 | Anthropic | Proprietary |
| 138 | Mistral Large 3MistralAI | 1,414 | +/-3 | 78,986 | MistralAI | Apache 2.0 |
| 139 | Haiku 4.5Anthropic | 1,414 | +/-2 | 142,743 | Anthropic | Proprietary |
| 140 | Qwen3-VL-235B-A22B-Instruct阿里巴巴 | 1,413 | +/-7 | 11,250 | 阿里巴巴 | Apache 2.0 |
| 141 | amazon-nova-experimental-chat-26-01-10Amazon | 1,413 | +/-10 | 3,534 | Amazon | Proprietary |
| 142 | Grok 3xAI | 1,411 | +/-4 | 32,389 | xAI | Proprietary |
| 143 | GLM-4.5智谱AI | 1,411 | +/-5 | 23,668 | 智谱AI | MIT |
| 144 | Grok 4xAI | 1,411 | +/-4 | 39,409 | xAI | Proprietary |
| 145 | Gemini 2.5 FlashGoogle DeepMind | 1,409 | +/-2 | 125,029 | Google DeepMind | Proprietary |
| 146 | hunyuan-hy3-previewTencent | 1,409 | +/-7 | 6,871 | Tencent | tencent-hunyuan-community |
| 147 | Qwen3.5-27B阿里巴巴 | 1,409 | +/-4 | 28,618 | 阿里巴巴 | Apache 2.0 |
| 148 | Magistral-Medium-2506MistralAI | 1,408 | +/-3 | 94,995 | MistralAI | Proprietary |
| 149 | Inkling SmallThinky | 1,405 | +/-5 | 26,446 | Thinky | Apache 2.0 |
| 150 | Grok 4 Fast (fast-reasoning)xAI | 1,405 | +/-5 | 18,148 | xAI | Proprietary |
| 151 | Gemini 2.5 Flash-Preview-09-2025Google DeepMind | 1,403 | +/-4 | 32,847 | Google DeepMind | Proprietary |
| 152 | qwen3-235b-a22b-no-thinkingAlibaba | 1,403 | +/-5 | 37,284 | Alibaba | Apache 2.0 |
| 153 | OpenAI o1OpenAI | 1,402 | +/-4 | 27,807 | OpenAI | Proprietary |
| 154 | Claude Sonnet 4 (thinking-32k)Anthropic | 1,402 | +/-4 | 33,923 | Anthropic | Proprietary |
| 155 | longcat-flash-chatMeituan | 1,401 | +/-6 | 11,123 | Meituan | MIT |
| 156 | GPT-5.4 nano (high)OpenAI | 1,401 | +/-4 | 62,300 | OpenAI | Proprietary |
| 157 | qwen3-235b-a22b-thinking-2507Alibaba | 1,400 | +/-7 | 8,776 | Alibaba | Apache 2.0 |
| 158 | Qwen3-Next阿里巴巴 | 1,399 | +/-5 | 22,607 | 阿里巴巴 | Apache 2.0 |
| 159 | DeepSeek-R1DeepSeek-AI | 1,398 | +/-5 | 18,524 | DeepSeek-AI | MIT |
| 160 |  Step 3.5 FlashStepFunAI | 1,396 | +/-4 | 61,687 | StepFunAI | Proprietary |
| 161 | DeepSeek-V3-0324DeepSeek-AI | 1,396 | +/-4 | 44,762 | DeepSeek-AI | MIT |
| 162 | hunyuan-vision-1.5-thinkingTencent | 1,395 | +/-12 | 2,159 | Tencent | Proprietary |
| 163 | Qwen3-VL-235B-A22B-Instruct (thinking)阿里巴巴 | 1,395 | +/-7 | 7,809 | 阿里巴巴 | Apache 2.0 |
| 164 | amazon-nova-experimental-chat-12-10Amazon | 1,394 | +/-9 | 3,727 | Amazon | Proprietary |
| 165 | Qwen3.5-35B-A3B阿里巴巴 | 1,394 | +/-4 | 30,410 | 阿里巴巴 | Apache 2.0 |
| 166 | Step 3.5 FlashStepFunAI | 1,393 | +/-4 | 60,236 | StepFunAI | Apache 2.0 |
| 167 | mimo-v2-flash (non-thinking)Xiaomi | 1,392 | +/-3 | 48,012 | Xiaomi | MIT |
| 168 | MiniMax M2.5MiniMaxAI | 1,391 | +/-4 | 42,734 | MiniMaxAI | Modified MIT |
| 169 | OpenAI o4 - miniOpenAI | 1,391 | +/-4 | 44,636 | OpenAI | Proprietary |
| 170 | Claude Sonnet 4Anthropic | 1,391 | +/-4 | 38,917 | Anthropic | Proprietary |
| 171 | GPT-5-mini (high)OpenAI | 1,390 | +/-5 | 26,645 | OpenAI | Proprietary |
| 172 | OpenAI o1OpenAI | 1,389 | +/-5 | 31,122 | OpenAI | Proprietary |
| 173 | Claude3-Sonnet (thinking-32k)Anthropic | 1,388 | +/-4 | 38,036 | Anthropic | Proprietary |
| 174 | mistral-medium-2505Mistral | 1,387 | +/-5 | 32,616 | Mistral | Proprietary |
| 175 | Qwen3-Coder-480B-A35B阿里巴巴 | 1,387 | +/-5 | 25,011 | 阿里巴巴 | Apache 2.0 |
| 176 | Hunyuan-T1腾讯AI实验室 | 1,387 | +/-9 | 4,574 | 腾讯AI实验室 | Proprietary |
| 177 | mimo-v2-flash (thinking)Xiaomi | 1,387 | +/-6 | 11,118 | Xiaomi | MIT |
| 178 | solar-pro4Upstage | 1,386 | +/-6 | 15,135 | Upstage | Proprietary |
| 179 | M2.1MiniMaxAI | 1,385 | +/-5 | 17,187 | MiniMaxAI | MIT |
| 180 | Qwen3-30B-A3B-2507阿里巴巴 | 1,383 | +/-5 | 23,182 | 阿里巴巴 | Apache 2.0 |
| 181 | GPT-4.1 miniOpenAI | 1,383 | +/-4 | 38,549 | OpenAI | Proprietary |
| 182 | hunyuan-turbos-20250416Tencent | 1,383 | +/-7 | 10,577 | Tencent | Proprietary |
| 183 | Gemini 2.5 Flash-Lite-Preview-09-2025 (no-thinking)Google DeepMind | 1,379 | +/-3 | 47,253 | Google DeepMind | Proprietary |
| 184 | GLM-4.6V智谱AI | 1,379 | +/-11 | 2,830 | 智谱AI | MIT |
| 185 | trinity-large-preview Apache 2.0 | 1,379 | +/-4 | 31,220 | — | — |
| 186 | Qwen3-235B-A22B阿里巴巴 | 1,375 | +/-5 | 25,826 | 阿里巴巴 | Apache 2.0 |
| 187 | Gemini 2.5 Flash-Lite (thinking)Google DeepMind | 1,375 | +/-5 | 32,178 | Google DeepMind | Proprietary |
| 188 | Claude 3.5 SonnetAnthropic | 1,374 | +/-3 | 87,644 | Anthropic | Proprietary |
| 189 | Qwen2.5-Max阿里巴巴 | 1,374 | +/-4 | 32,411 | 阿里巴巴 | Proprietary |
| 190 | GLM-4.5-Air智谱AI | 1,373 | +/-4 | 30,454 | 智谱AI | MIT |
| 191 | Claude3-SonnetAnthropic | 1,372 | +/-4 | 42,411 | Anthropic | Proprietary |
| 192 | Qwen3-Next (thinking)阿里巴巴 | 1,369 | +/-6 | 13,382 | 阿里巴巴 | Apache 2.0 |
| 193 | trinity-large-thinking Apache 2.0 | 1,367 | +/-5 | 30,281 | — | — |
| 194 | Gemma 3 - 27B (IT)Google DeepMind | 1,365 | +/-4 | 46,437 | Google DeepMind | Gemma |
| 195 | GLM-4.7-Flash智谱AI | 1,365 | +/-6 | 11,979 | 智谱AI | MIT |
| 196 | amazon-nova-experimental-chat-11-10Amazon | 1,364 | +/-4 | 25,992 | Amazon | Proprietary |
| 197 | minimax-m1MiniMax | 1,364 | +/-4 | 34,048 | MiniMax | Apache 2.0 |
| 198 | OpenAI o3-mini (high)OpenAI | 1,364 | +/-5 | 18,589 | OpenAI | Proprietary |
| 199 | OpenAI o3-mini (high)OpenAI | 1,363 | +/-5 | 15,506 | OpenAI | Proprietary |
| 200 | nvidia-nemotron-3-super-120b-a12bNvidia | 1,361 | +/-7 | 7,861 | Nvidia | NVIDIA Open Model |
| 201 | Gemini 2.0 Flash ExperimentalGoogle DeepMind | 1,360 | +/-4 | 43,317 | Google DeepMind | Proprietary |
| 202 | DeepSeek-V3DeepSeek-AI | 1,358 | +/-5 | 21,770 | DeepSeek-AI | DeepSeek |
| 203 | grok-3-mini-betaxAI | 1,358 | +/-5 | 20,814 | xAI | Proprietary |
| 204 | Mistral-Small-3.2MistralAI | 1,356 | +/-5 | 17,353 | MistralAI | Apache 2.0 |
| 205 | intellect-3 MIT | 1,356 | +/-8 | 5,421 | — | — |
| 206 |  C4AI Command A (202503)CohereAI | 1,354 | +/-3 | 55,477 | CohereAI | CC-BY-NC-4.0 |
| 207 | Gemini 2.0 Flash-LiteGoogle DeepMind | 1,354 | +/-4 | 24,955 | Google DeepMind | Proprietary |
| 208 | GLM-4.5V智谱AI | 1,352 | +/-8 | 4,811 | 智谱AI | MIT |
| 209 | GPT OSS 120BOpenAI | 1,352 | +/-4 | 30,039 | OpenAI | Apache 2.0 |
| 210 | Gemini 1.5 ProGoogle DeepMind | 1,351 | +/-3 | 55,606 | Google DeepMind | Proprietary |
| 211 | Step3StepFunAI | 1,349 | +/-8 | 6,272 | StepFunAI | Apache 2.0 |
| 212 | hunyuan-turbos-20250226Tencent | 1,349 | +/-12 | 2,220 | Tencent | Proprietary |
| 213 | nvidia-nemotron-3.5-lightning-30b-a3b-nvfp4Nvidia | 1,349 | +/-6 | 14,326 | Nvidia | OpenMDW-1.1 |
| 214 | amazon-nova-experimental-chat-10-20Amazon | 1,348 | +/-6 | 11,596 | Amazon | Proprietary |
| 215 | OpenAI o3-miniOpenAI | 1,348 | +/-4 | 56,573 | OpenAI | Proprietary |
| 216 | llama-3.1-nemotron-ultra-253b-v1Nvidia | 1,348 | +/-12 | 2,549 | Nvidia | Nvidia Open Model |
| 217 | amazon-nova-experimental-chat-10-09Amazon | 1,347 | +/-11 | 2,765 | Amazon | Proprietary |
| 218 | Qwen3-32B阿里巴巴 | 1,347 | +/-9 | 3,926 | 阿里巴巴 | Apache 2.0 |
| 219 | GPT-4oOpenAI | 1,346 | +/-3 | 112,881 | OpenAI | Proprietary |
| 220 | qwen-plus-0125Alibaba | 1,346 | +/-8 | 5,819 | Alibaba | Proprietary |
| 221 | ling-flash-2.0 AntGroup | 1,344 | +/-7 | 6,889 | Group | MIT |
| 222 | MiniMax M2MiniMaxAI | 1,343 | +/-8 | 6,819 | MiniMaxAI | Apache 2.0 |
| 223 | mercury-2 InceptionAI | 1,343 | +/-11 | 3,198 | AI | Proprietary |
| 224 | Claude 3.5 SonnetAnthropic | 1,343 | +/-3 | 82,419 | Anthropic | Proprietary |
| 225 | nvidia-llama-3.3-nemotron-super-49b-v1.5Nvidia | 1,343 | +/-10 | 3,308 | Nvidia | Nvidia Open |
| 226 | glm-4-plus-0111Zhipu | 1,343 | +/-8 | 5,760 | Zhipu | Proprietary |
| 227 | Gemma 3 - 12B (IT)Google DeepMind | 1,342 | +/-10 | 3,829 | Google DeepMind | Gemma |
| 228 | hunyuan-turbo-0110Tencent | 1,341 | +/-12 | 2,290 | Tencent | Proprietary |
| 229 | granite-4.2-30bIBM | 1,340 | +/-10 | 3,416 | IBM | Apache 2.0 |
| 230 | GPT-5-Nano (high)OpenAI | 1,337 | +/-7 | 8,115 | OpenAI | Proprietary |
| 231 | OpenAI o1-miniOpenAI | 1,337 | +/-4 | 51,981 | OpenAI | Proprietary |
| 232 | gemini-advanced-0514Google | 1,336 | +/-5 | 50,148 | Proprietary | |
| 233 | QwQ-32B阿里巴巴 | 1,336 | +/-4 | 25,084 | 阿里巴巴 | Apache 2.0 |
| 234 | Grok 2xAI | 1,336 | +/-4 | 63,498 | xAI | Proprietary |
| 235 | GPT-4oOpenAI | 1,336 | +/-4 | 45,499 | OpenAI | Proprietary |
| 236 | llama-3.1-405b-instruct-bf16Meta | 1,335 | +/-4 | 41,375 | Meta | Llama 3.1 Community |
| 237 | Nova 2 Lite亚马逊 | 1,335 | +/-6 | 12,393 | 亚马逊 | Proprietary |
| 238 | step-2-16k-exp-202412StepFun | 1,334 | +/-9 | 4,833 | StepFun | Proprietary |
| 239 | llama-3.1-405b-instruct-fp8Meta | 1,333 | +/-4 | 59,656 | Meta | Llama 3.1 Community |
| 240 | olmo-3.1-32b-instructAi2 | 1,329 | +/-6 | 11,794 | Ai2 | Apache 2.0 |
| 241 | yi-lightning Proprietary | 1,328 | +/-5 | 27,332 | — | — |
| 242 | llama-3.3-nemotron-49b-super-v1Nvidia | 1,328 | +/-12 | 2,218 | Nvidia | Nvidia |
| 243 | Llama 4 Maverick InstructFacebook AI研究实验室 | 1,327 | +/-4 | 39,266 | Facebook AI研究实验室 | Llama 4 |
| 244 | Qwen3-30B-A3B阿里巴巴 | 1,327 | +/-5 | 26,036 | 阿里巴巴 | Apache 2.0 |
| 245 | hunyuan-large-2025-02-10Tencent | 1,326 | +/-10 | 3,738 | Tencent | Proprietary |
| 246 | Claude 3.5 HaikuAnthropic | 1,325 | +/-3 | 68,877 | Anthropic | Proprietary |
| 247 | gpt-4-turbo-2024-04-09OpenAI | 1,324 | +/-4 | 98,114 | OpenAI | Proprietary |
| 248 | Gemini 1.5 ProGoogle DeepMind | 1,324 | +/-4 | 79,138 | Google DeepMind | Proprietary |
| 249 | deepseek-v2.5-1210DeepSeek | 1,323 | +/-8 | 6,795 | DeepSeek | DeepSeek |
| 250 | ring-flash-2.0 AntGroup | 1,323 | +/-7 | 6,934 | Group | MIT |
| 251 | GPT-4.1 nanoOpenAI | 1,322 | +/-8 | 6,103 | OpenAI | Proprietary |
| 252 | Claude3-OpusAnthropic | 1,322 | +/-3 | 194,909 | Anthropic | Proprietary |
| 253 | molmo-2-8bAi2 | 1,322 | +/-21 | 818 | Ai2 | Apache 2.0 |
| 254 | Llama 4 Scout InstructFacebook AI研究实验室 | 1,322 | +/-5 | 29,668 | Facebook AI研究实验室 | Llama |
| 255 | step-1o-turbo-202506StepFun | 1,320 | +/-7 | 8,845 | StepFun | Proprietary |
| 256 | glm-4-plusZhipu AI | 1,319 | +/-5 | 26,126 | Zhipu AI | Proprietary |
| 257 | qwen-max-0919Alibaba | 1,318 | +/-6 | 16,478 | Alibaba | Qwen |
| 258 | Llama3.3-70B-InstructFacebook AI研究实验室 | 1,318 | +/-4 | 54,371 | Facebook AI研究实验室 | Llama-3.3 |
| 259 | GPT-4o miniOpenAI | 1,318 | +/-4 | 68,697 | OpenAI | Proprietary |
| 260 | GPT OSS 20BOpenAI | 1,318 | +/-6 | 10,371 | OpenAI | Apache 2.0 |
| 261 | Gemma-3n-E4BGoogle DeepMind | 1,317 | +/-5 | 22,038 | Google DeepMind | Gemma |
| 262 | qwen2.5-plus-1127Alibaba | 1,315 | +/-6 | 10,187 | Alibaba | Proprietary |
| 263 | mistral-large-2407Mistral | 1,314 | +/-4 | 45,459 | Mistral | Mistral Research |
| 264 | athene-v2-chat NexusFlow | 1,314 | +/-5 | 24,739 | — | — |
| 265 | nvidia-nemotron-3-nano-30b-a3b-bf16Nvidia | 1,314 | +/-5 | 15,724 | Nvidia | NVIDIA Open Model |
| 266 | GPT-4OpenAI | 1,313 | +/-4 | 93,439 | OpenAI | Proprietary |
| 267 | GPT-4OpenAI | 1,313 | +/-4 | 100,105 | OpenAI | Proprietary |
| 268 | hunyuan-standard-2025-02-10Tencent | 1,311 | +/-10 | 3,904 | Tencent | Proprietary |
| 269 | gemini-1.5-flash-002Google | 1,309 | +/-4 | 34,902 | Proprietary | |
| 270 | grok-2-mini-2024-08-13xAI | 1,308 | +/-4 | 52,567 | xAI | Proprietary |
| 271 | DeepSeek V2.5DeepSeek-AI | 1,307 | +/-5 | 24,572 | DeepSeek-AI | DeepSeek |
| 272 | athene-70b-0725 CC-BY-NC-4.0 | 1,307 | +/-6 | 19,621 | — | — |
| 273 | olmo-3-32b-thinkAi2 | 1,306 | +/-8 | 5,811 | Ai2 | Apache 2.0 |
| 274 | mercury InceptionAI | 1,306 | +/-14 | 1,948 | AI | Proprietary |
| 275 | mistral-large-2411Mistral | 1,306 | +/-4 | 28,073 | Mistral | MRL |
| 276 | Magistral-Medium-2506MistralAI | 1,305 | +/-6 | 11,435 | MistralAI | Proprietary |
| 277 | granite-4.1-8bIBM | 1,305 | +/-10 | 4,166 | IBM | Apache 2.0 |
| 278 | Gemma 3 - 4B (IT)Google DeepMind | 1,303 | +/-9 | 4,171 | Google DeepMind | Gemma |
| 279 | Mistral-Small-3.1-24B-Instruct-2503MistralAI | 1,303 | +/-5 | 32,556 | MistralAI | Apache 2.0 |
| 280 | Qwen2.5-VL-72B-Instruct阿里巴巴 | 1,303 | +/-4 | 39,406 | 阿里巴巴 | Qwen |
| 281 | Llama3.1-70B-InstructFacebook AI研究实验室 | 1,299 | +/-8 | 7,140 | Facebook AI研究实验室 | Llama 3.1 |
| 282 | hunyuan-large-visionTencent | 1,294 | +/-9 | 5,250 | Tencent | Proprietary |
| 283 | Llama3.1-70B-InstructFacebook AI研究实验室 | 1,293 | +/-4 | 55,240 | Facebook AI研究实验室 | Llama 3.1 Community |
| 284 | amazon-nova-pro-v1.0Amazon | 1,290 | +/-5 | 24,745 | Amazon | Proprietary |
| 285 | gemma-2-27b-itGoogle | 1,289 | +/-3 | 75,754 | Gemma license | |
| 286 | jamba-1.5-large Jamba Open | 1,289 | +/-7 | 8,662 | — | — |
| 287 | granite-4.2-3bIBM | 1,289 | +/-12 | 3,266 | IBM | Apache 2.0 |
| 288 | reka-core-20240904 Proprietary | 1,288 | +/-7 | 7,312 | — | — |
| 289 | GPT-4OpenAI | 1,288 | +/-5 | 54,173 | OpenAI | Proprietary |
| 290 | granite-4.2-8bIBM | 1,288 | +/-11 | 3,255 | IBM | Apache 2.0 |
| 291 | ibm-granite-h-smallIBM | 1,287 | +/-8 | 5,581 | IBM | Apache 2.0 |
| 292 | llama-3.1-nemotron-51b-instructNvidia | 1,287 | +/-10 | 3,749 | Nvidia | Llama 3.1 |
| 293 | gemini-1.5-flash-001Google | 1,287 | +/-5 | 62,833 | Proprietary | |
| 294 | olmo-3.1-32b-thinkAi2 | 1,286 | +/-7 | 8,278 | Ai2 | Apache 2.0 |
| 295 | llama-3.1-tulu-3-70bAi2 | 1,286 | +/-10 | 2,846 | Ai2 | Llama 3.1 |
| 296 | Claude3-SonnetAnthropic | 1,281 | +/-4 | 109,284 | Anthropic | Proprietary |
| 297 | gemma-2-9b-it-simpo MIT | 1,280 | +/-7 | 10,072 | — | — |
| 298 | nemotron-4-340b-instructNvidia | 1,277 | +/-5 | 19,659 | Nvidia | NVIDIA Open Model |
| 299 | Llama3-70B-InstructFacebook AI研究实验室 | 1,276 | +/-4 | 156,876 | Facebook AI研究实验室 | Llama 3 Community |
| 300 | command-r-plus-08-2024Cohere | 1,276 | +/-7 | 9,866 | Cohere | CC-BY-NC-4.0 |
| 301 | GPT-4OpenAI | 1,276 | +/-4 | 88,723 | OpenAI | Proprietary |
| 302 | Mistral Small 24B Instruct 2501MistralAI | 1,274 | +/-6 | 14,681 | MistralAI | Apache 2.0 |
| 303 | GLM4智谱AI | 1,273 | +/-7 | 9,788 | 智谱AI | Proprietary |
| 304 | reka-flash-20240904 Proprietary | 1,272 | +/-7 | 7,536 | — | — |
| 305 | Qwen2.5-Coder-32B-Instruct阿里巴巴 | 1,271 | +/-8 | 5,432 | 阿里巴巴 | Apache 2.0 |
| 306 | C4AI Aya Vision 32BCohereAI | 1,267 | +/-5 | 27,124 | CohereAI | CC-BY-NC-4.0 |
| 307 | gemma-2-9b-itGoogle | 1,267 | +/-4 | 54,611 | Gemma license | |
| 308 | deepseek-coder-v2DeepSeek | 1,265 | +/-6 | 15,147 | DeepSeek | DeepSeek License |
| 309 | Qwen2-72B-Instruct阿里巴巴 | 1,262 | +/-5 | 37,325 | 阿里巴巴 | Qianwen LICENSE |
| 310 | C4AI Command R+CohereAI | 1,262 | +/-4 | 77,554 | CohereAI | CC-BY-NC-4.0 |
| 311 | Claude3-HaikuAnthropic | 1,262 | +/-4 | 117,701 | Anthropic | Proprietary |
| 312 | amazon-nova-lite-v1.0Amazon | 1,260 | +/-5 | 19,372 | Amazon | Proprietary |
| 313 | gemini-1.5-flash-8b-001Google | 1,259 | +/-4 | 35,558 | Proprietary | |
| 314 |  Phi-4-reasoningMicrosoft Azure | 1,256 | +/-5 | 24,126 | Microsoft Azure | MIT |
| 315 | olmo-2-0325-32b-instructAi2 | 1,252 | +/-11 | 3,334 | Ai2 | Apache-2.0 |
| 316 | command-r-08-2024Cohere | 1,250 | +/-7 | 10,140 | Cohere | CC-BY-NC-4.0 |
| 317 | mistral-large-2402Mistral | 1,242 | +/-5 | 62,436 | Mistral | Proprietary |
| 318 | amazon-nova-micro-v1.0Amazon | 1,241 | +/-5 | 19,364 | Amazon | Proprietary |
| 319 | jamba-1.5-mini Jamba Open | 1,240 | +/-7 | 8,858 | — | — |
| 320 | ministral-8b-2410Mistral | 1,238 | +/-9 | 4,781 | Mistral | MRL |
| 321 | gemini-pro-dev-apiGoogle | 1,237 | +/-7 | 18,354 | Proprietary | |
| 322 | Qwen1.5-110B-Chat阿里巴巴 | 1,234 | +/-6 | 26,195 | 阿里巴巴 | Qianwen LICENSE |
| 323 | reka-flash-21b-20240226-online Proprietary | 1,233 | +/-7 | 15,450 | — | — |
| 324 | Qwen1.5-72B-Chat阿里巴巴 | 1,233 | +/-5 | 39,302 | 阿里巴巴 | Qianwen LICENSE |
| 325 | hunyuan-standard-256kTencent | 1,233 | +/-12 | 2,728 | Tencent | Proprietary |
| 326 | Mixtral-8x22B-Instruct-v0.1MistralAI | 1,230 | +/-5 | 51,416 | MistralAI | Apache 2.0 |
| 327 | command-rCohere | 1,227 | +/-5 | 54,036 | Cohere | CC-BY-NC-4.0 |
| 328 | reka-flash-21b-20240226 Proprietary | 1,227 | +/-6 | 24,806 | — | — |
| 329 | gpt-3.5-turbo-0125OpenAI | 1,226 | +/-5 | 66,207 | OpenAI | Proprietary |
| 330 | Llama3-8B-InstructFacebook AI研究实验室 | 1,224 | +/-4 | 104,642 | Facebook AI研究实验室 | Llama 3 Community |
| 331 | Gemini-proGoogle DeepMind | 1,224 | +/-12 | 6,390 | Google DeepMind | Proprietary |
| 332 | C4AI Aya Vision 8BCohereAI | 1,223 | +/-7 | 9,818 | CohereAI | CC-BY-NC-4.0 |
| 333 | mistral-mediumMistral | 1,223 | +/-5 | 34,550 | Mistral | Proprietary |
| 334 | llama-3.1-tulu-3-8bAi2 | 1,220 | +/-11 | 2,896 | Ai2 | Llama 3.1 |
| 335 | zephyr-orpo-141b-A35b-v0.1 Apache 2.0 | 1,213 | +/-11 | 4,652 | — | — |
| 336 |  Yi-1.5-34B零一万物 | 1,213 | +/-5 | 24,146 | 零一万物 | — |
| 337 | Llama3.1-8B-InstructFacebook AI研究实验室 | 1,211 | +/-4 | 49,605 | Facebook AI研究实验室 | Llama 3.1 Community |
| 338 | Llama3.1-8B-InstructFacebook AI研究实验室 | 1,208 | +/-11 | 3,090 | Facebook AI研究实验室 | Apache 2.0 |
| 339 | gpt-3.5-turbo-1106OpenAI | 1,204 | +/-9 | 16,619 | OpenAI | Proprietary |
| 340 | qwen1.5-32b-chatAlibaba | 1,204 | +/-6 | 21,741 | Alibaba | Qianwen LICENSE |
| 341 | gemma-2-2b-itGoogle | 1,200 | +/-4 | 46,616 | Gemma license | |
| 342 | Phi-3-medium 14B-previewMicrosoft Azure | 1,198 | +/-5 | 25,055 | Microsoft Azure | MIT |
| 343 | mixtral-8x7b-instruct-v0.1Mistral | 1,197 | +/-4 | 73,503 | Mistral | Apache 2.0 |
| 344 | dbrx-instruct-preview DBRX LICENSE | 1,195 | +/-6 | 32,191 | — | — |
| 345 | Qwen1.5-14B-Chat阿里巴巴 | 1,191 | +/-7 | 17,839 | 阿里巴巴 | Qianwen LICENSE |
| 346 |  InternLM2-Base-20B上海人工智能实验室 | 1,191 | +/-7 | 9,901 | 上海人工智能实验室 | — |
| 347 | DeepSeek LLM 67B ChatDeepSeek-AI | 1,185 | +/-11 | 4,932 | DeepSeek-AI | DeepSeek License |
| 348 |  WizardLM-70B-V1.0WizardLM Team | 1,185 | +/-9 | 8,214 | WizardLM Team | Llama 2 Community |
| 349 | Yi-34B零一万物 | 1,184 | +/-7 | 15,483 | 零一万物 | — |
| 350 | granite-3.0-8b-instructIBM | 1,183 | +/-9 | 6,638 | IBM | Apache 2.0 |
| 351 | openchat-3.5 Apache-2.0 | 1,183 | +/-10 | 7,968 | — | — |
| 352 | openchat-3.5-0106 Apache-2.0 | 1,183 | +/-8 | 12,637 | — | — |
| 353 |  Gemma 1.1-7B-ITGoogle Research | 1,183 | +/-6 | 23,893 | Google Research | Gemma license |
| 354 | snowflake-arctic-instruct Apache 2.0 | 1,180 | +/-6 | 32,832 | — | — |
| 355 | granite-3.1-2b-instructIBM | 1,179 | +/-11 | 3,188 | IBM | Apache 2.0 |
| 356 | LLaMA2 70BFacebook AI研究实验室 | 1,178 | +/-10 | 6,535 | Facebook AI研究实验室 | — |
| 357 | openhermes-2.5-mistral-7b Apache-2.0 | 1,176 | +/-10 | 5,006 | — | — |
| 358 |  Vicuna 33BLM-SYS | 1,173 | +/-6 | 22,479 | LM-SYS | — |
| 359 | Phi-3-small 7BMicrosoft Azure | 1,171 | +/-6 | 17,766 | Microsoft Azure | MIT |
| 360 | starling-lm-7b-beta Apache-2.0 | 1,171 | +/-7 | 16,056 | — | — |
| 361 | llama-2-70b-chatMeta | 1,171 | +/-5 | 38,492 | Meta | Llama 2 Community |
| 362 | starling-lm-7b-alpha CC-BY-NC-4.0 | 1,167 | +/-8 | 10,224 | — | — |
| 363 | llama-3.2-3b-instructMeta | 1,167 | +/-8 | 7,936 | Meta | Llama 3.2 |
| 364 | nous-hermes-2-mixtral-8x7b-dpo Apache-2.0 | 1,164 | +/-12 | 3,777 | — | — |
| 365 | Qwen3-VL-2B阿里巴巴 | 1,157 | +/-8 | 6,837 | 阿里巴巴 | Apache 2.0 |
| 366 | llama2-70b-steerlm-chatNvidia | 1,155 | +/-13 | 3,585 | Nvidia | Llama 2 Community |
| 367 | QwQ-32B-Preview阿里巴巴 | 1,154 | +/-11 | 3,231 | 阿里巴巴 | Apache 2.0 |
| 368 | solar-10.7b-instruct-v1.0 CC-BY-NC-4.0 | 1,152 | +/-13 | 4,155 | — | — |
| 369 | dolphin-2.2.1-mistral-7b Apache-2.0 | 1,152 | +/-15 | 1,679 | — | — |
| 370 | mpt-30b-chat CC-BY-NC-SA-4.0 | 1,151 | +/-12 | 2,572 | — | — |
| 371 | wizardlm-13bMicrosoft | 1,149 | +/-9 | 7,044 | Microsoft | Llama 2 Community |
| 372 | Mistral-7B-Instruct-v0.2MistralAI | 1,149 | +/-7 | 19,402 | MistralAI | Apache-2.0 |
| 373 | falcon-180b-chat Falcon-180B TII License | 1,148 | +/-17 | 1,295 | — | — |
| 374 | Qwen1.5-7B-Chat阿里巴巴 | 1,144 | +/-10 | 4,737 | 阿里巴巴 | Qianwen LICENSE |
| 375 | Phi-3-mini 3.8BMicrosoft Azure | 1,143 | +/-6 | 12,297 | Microsoft Azure | MIT |
| 376 | vicuna-13b Llama 2 Community | 1,141 | +/-7 | 19,367 | — | — |
| 377 |  Baichuan2-13B-Chat百川智能 | 1,141 | +/-7 | 19,174 | 百川智能 | Llama 2 Community |
| 378 | Qwen-14B-Chat阿里巴巴 | 1,139 | +/-11 | 4,964 | 阿里巴巴 | Qianwen LICENSE |
| 379 | PaLM 2Google Research | 1,139 | +/-9 | 8,554 | Google Research | Proprietary |
| 380 | Gemma 7B - ItGoogle Research | 1,138 | +/-9 | 8,925 | Google Research | Gemma license |
| 381 | CodeLLaMA-34BFacebook AI研究实验室 | 1,137 | +/-9 | 7,366 | Facebook AI研究实验室 | Llama 2 Community |
| 382 | zephyr-7b-beta MIT | 1,131 | +/-9 | 11,118 | — | — |
| 383 | Phi-3-mini 3.8BMicrosoft Azure | 1,130 | +/-7 | 20,685 | Microsoft Azure | MIT |
| 384 | Phi-3-mini 3.8BMicrosoft Azure | 1,128 | +/-6 | 20,118 | Microsoft Azure | MIT |
| 385 | guanaco-33b Non-commercial | 1,127 | +/-12 | 2,921 | — | — |
| 386 | zephyr-7b-alpha MIT | 1,127 | +/-16 | 1,785 | — | — |
| 387 | stripedhyena-nous-7b Apache 2.0 | 1,122 | +/-11 | 5,182 | — | — |
| 388 | CodeLlama-70B-InstructFacebook AI研究实验室 | 1,119 | +/-18 | 1,143 | Facebook AI研究实验室 | Llama 2 Community |
| 389 | Gemma 1.1-2B-ITGoogle Research | 1,117 | +/-8 | 10,854 | Google Research | Gemma license |
| 390 | vicuna-7b Llama 2 Community | 1,115 | +/-9 | 6,923 | — | — |
| 391 | smollm2-1.7b-instruct Apache 2.0 | 1,115 | +/-14 | 2,199 | — | — |
| 392 | llama-3.2-1b-instructMeta | 1,111 | +/-8 | 8,045 | Meta | Llama 3.2 |
| 393 | Mistral 7B InstructMistralAI | 1,110 | +/-9 | 8,977 | MistralAI | Apache 2.0 |
| 394 | Baichuan2-7B-Chat百川智能 | 1,108 | +/-7 | 14,148 | 百川智能 | Llama 2 Community |
| 395 | Gemma 2B - ItGoogle Research | 1,094 | +/-11 | 4,780 | Google Research | Gemma license |
| 396 | Qwen1.5-4B-Chat阿里巴巴 | 1,091 | +/-9 | 7,597 | 阿里巴巴 | Qianwen LICENSE |
| 397 | olmo-7b-instructAi2 | 1,074 | +/-11 | 6,328 | Ai2 | Apache-2.0 |
| 398 |  Koala达摩院 | 1,071 | +/-10 | 6,965 | 达摩院 | — |
| 399 | alpaca-13b Non-commercial | 1,070 | +/-11 | 5,745 | — | — |
| 400 |  GPT4All 13BNomic AI | 1,068 | +/-15 | 1,743 | Nomic AI | — |
| 401 | mpt-7b-chat CC-BY-NC-SA-4.0 | 1,063 | +/-12 | 3,924 | — | — |
| 402 | ChatGLM3-6B智谱AI | 1,056 | +/-12 | 4,658 | 智谱AI | — |
| 403 | RWKV-4-Raven-14B Apache 2.0 | 1,042 | +/-11 | 4,845 | — | — |
| 404 | ChatGLM2-6B智谱AI | 1,025 | +/-14 | 2,658 | 智谱AI | — |
| 405 | oasst-pythia-12b Apache 2.0 | 1,023 | +/-11 | 6,310 | — | — |
| 406 | ChatGLM-6B智谱AI | 996 | +/-13 | 4,914 | 智谱AI | — |
| 407 | fastchat-t5-3b Apache 2.0 | 992 | +/-12 | 4,203 | — | — |
| 408 | dolly-v2-12b MIT | 982 | +/-13 | 3,412 | — | — |
| 409 | LLaMA 13BFacebook AI研究实验室 | 975 | +/-16 | 2,391 | Facebook AI研究实验室 | Non-commercial |
| 410 | stablelm-tuned-alpha-7b CC-BY-NC-SA-4.0 | 953 | +/-13 | 3,287 | — | — |
数据仅供参考，以官方来源为准。模型名称旁的链接可跳转到 DataLearner 模型详情页。
常见问题 (FAQ)
01
什么是 Text Generation Arena (LMArena)？
Text Generation Arena（原 LMSYS Chatbot Arena）是目前最具影响力的大模型匿名评测平台。用户向两个身份未知的模型提问，根据回答质量投票，系统通过 Elo 算法将数百万次投票汇聚为动态排行榜，被学术界和工业界广泛引用。
02
Arena Elo 分数是如何计算的？
Elo 算法源自国际象棋评分体系。每次对战后，胜者得分上升、败者下降，幅度取决于双方原始评分差距。95% 置信区间（CI）反映该模型参与对战次数的多少：CI 越窄说明数据越充分、排名越可信。
03
为什么同一模型会出现"Thinking"和普通两个版本？
部分模型支持"扩展思考"（Extended Thinking）模式，会在给出最终答案前进行更深入的内部推理。该模式通常在逻辑推理、数学和编程任务上得分更高，但响应时延也更长、成本更高。Arena 将两种模式分开评测，以便用户根据实际需求选择。
04
如何根据排行榜选择适合自己的大语言模型？
建议综合考虑：综合性能（看 Elo 总分）、成本（闭源 API 按量计费，开源可自部署）、中文支持、开源程度以及响应速度。