# lmarena.ai排行榜 > 排名更新时间: 2025-01-23 | 排名(UB) | 排名(StyleCtrl) | 模型名 | 分数 | 置信区间 | 票数 | 服务商 | | -----: | ------------: | ----------------------------------- | ---: | ------- | ------: | ---------------------- | | 1 | 3 | Gemini-2.0-Flash-Thinking-Exp-01-21 | 1382 | +8/-6 | 6,437 | Google | | 1 | 1 | Gemini-Exp-1206 | 1374 | +5/-4 | 22,116 | Google | | 3 | 1 | ChatGPT-4o-latest (2024-11-20) | 1365 | +4/-4 | 35,328 | OpenAI | | 3 | 1 | DeepSeek-R1 | 1357 | +12/-13 | 1,883 | DeepSeek | | 4 | 5 | Gemini-2.0-Flash-Exp | 1356 | +4/-4 | 20,939 | Google | | 4 | 1 | o1-2024-12-17 | 1352 | +6/-6 | 9,230 | OpenAI | | 7 | 4 | o1-preview | 1335 | +3/-3 | 33,186 | OpenAI | | 8 | 9 | DeepSeek-V3 | 1317 | +6/-5 | 13,640 | DeepSeek | | 8 | 11 | Step-2-16K-Exp | 1305 | +9/-7 | 4,533 | StepFun | | 9 | 12 | o1-mini | 1305 | +2/-3 | 49,952 | OpenAI | | 9 | 9 | Gemini-1.5-Pro-002 | 1302 | +3/-4 | 46,621 | Google | | 12 | 14 | Grok-2-08-13 | 1288 | +3/-3 | 67,150 | xAI | | 12 | 17 | Yi-Lightning | 1287 | +3/-4 | 28,955 | 01 AI | | 12 | 10 | GPT-4o-2024-05-13 | 1285 | +2/-2 | 117,745 | OpenAI | | 12 | 8 | Claude 3.5 Sonnet (20241022) | 1283 | +3/-3 | 48,847 | Anthropic | | 12 | 22 | Qwen2.5-plus-1127 | 1283 | +5/-7 | 9,050 | Alibaba | | 13 | 20 | Deepseek-v2.5-1210 | 1279 | +5/-6 | 7,261 | DeepSeek | | 15 | 24 | Athene-v2-Chat-72B | 1276 | +4/-5 | 22,355 | NexusFlow | | 16 | 22 | GLM-4-Plus | 1274 | +4/-4 | 27,771 | Zhipu AI | | 16 | 23 | GPT-4o-mini-2024-07-18 | 1273 | +3/-3 | 61,233 | OpenAI | | 17 | 24 | Gemini-1.5-Flash-002 | 1271 | +4/-3 | 35,199 | Google | | 17 | 36 | Llama-3.1-Nemotron-70B-Instruct | 1269 | +6/-5 | 7,598 | Nvidia | | 18 | 12 | Meta-Llama-3.1-405B-Instruct-bf16 | 1268 | +4/-4 | 22,703 | Meta | | 19 | 11 | Claude 3.5 Sonnet (20240620) | 1268 | +2/-3 | 86,167 | Anthropic | | 19 | 13 | Meta-Llama-3.1-405B-Instruct-fp8 | 1267 | +3/-3 | 63,187 | Meta | | 21 | 12 | Gemini Advanced App (2024-05-14) | 1267 | +3/-2 | 52,145 | Google | | 21 | 33 | Grok-2-Mini-08-13 | 1266 | +3/-3 | 55,507 | xAI | | 21 | 14 | GPT-4o-2024-08-06 | 1265 | +3/-3 | 47,975 | OpenAI | | 21 | 24 | Qwen-Max-0919 | 1263 | +6/-5 | 17,434 | Alibaba | | 29 | 20 | Gemini-1.5-Pro-001 | 1260 | +3/-2 | 82,433 | Google | | 29 | 29 | Deepseek-v2.5 | 1258 | +4/-4 | 26,345 | DeepSeek | | 29 | 36 | Qwen2.5-72B-Instruct | 1257 | +4/-3 | 40,664 | Alibaba | | 29 | 20 | GPT-4-Turbo-2024-04-09 | 1256 | +2/-2 | 102,125 | OpenAI | | 29 | 21 | Llama-3.3-70B-Instruct | 1256 | +4/-6 | 16,905 | Meta | | 34 | 24 | Mistral-Large-2407 | 1251 | +3/-3 | 48,205 | Mistral | | 34 | 34 | Athene-70B | 1250 | +3/-5 | 20,609 | NexusFlow | | 34 | 23 | GPT-4-1106-preview | 1250 | +2/-2 | 103,732 | OpenAI | | 34 | 40 | Meta-Llama-3.1-70B-Instruct | 1248 | +3/-3 | 58,785 | Meta | | 34 | 38 | Llama-3.1-Tulu-3-70B | 1244 | +9/-10 | 3,031 | Ai2 | | 35 | 21 | Claude 3 Opus | 1247 | +2/-2 | 202,713 | Anthropic | | 35 | 40 | Amazon Nova Pro 1.0 | 1243 | +6/-4 | 13,738 | Amazon | | 36 | 36 | Mistral-Large-2411 | 1244 | +5/-5 | 12,008 | Mistral | | 37 | 25 | GPT-4-0125-preview | 1245 | +2/-2 | 97,064 | OpenAI | | 40 | 21 | Claude 3.5 Haiku (20241022) | 1239 | +6/-7 | 12,181 | Anthropic | | 41 | 40 | Reka-Core-20240904 | 1235 | +6/-5 | 7,942 | Reka AI | | 46 | 43 | Gemini-1.5-Flash-001 | 1227 | +3/-2 | 65,656 | Google | | 46 | 40 | Jamba-1.5-Large | 1221 | +5/-5 | 9,125 | AI21 Labs | | 46 | 48 | Qwen2.5-Coder-32B-Instruct | 1217 | +8/-8 | 5,730 | Alibaba | | 47 | 43 | Gemma-2-27B-it | 1220 | +3/-3 | 73,168 | Google | | 47 | 56 | Amazon Nova Lite 1.0 | 1219 | +5/-6 | 11,563 | Amazon | | 47 | 44 | Command R+ (08-2024) | 1215 | +5/-5 | 10,541 | Cohere | | 47 | 43 | Gemma-2-9B-it-SimPO | 1216 | +5/-6 | 10,551 | Princeton | | 47 | 56 | Phi-4 | 1211 | +7/-10 | 4,039 | Microsoft | | 47 | 40 | Llama-3.1-Nemotron-51B-Instruct | 1211 | +8/-10 | 3,898 | Nvidia | | 49 | 58 | Gemini-1.5-Flash-8B-001 | 1212 | +4/-3 | 36,856 | Google | | 49 | 47 | Nemotron-4-340B-Instruct | 1209 | +4/-3 | 20,604 | Nvidia | | 49 | 56 | Aya-Expanse-32B | 1209 | +4/-4 | 27,764 | Cohere | | 50 | 52 | GLM-4-0520 | 1207 | +5/-5 | 10,212 | Zhipu AI | | 50 | 48 | Reka-Flash-20240904 | 1205 | +6/-7 | 8,131 | Reka AI | | 53 | 49 | Llama-3-70B-Instruct | 1206 | +2/-2 | 163,809 | Meta | | 57 | 48 | Claude 3 Sonnet | 1201 | +2/-3 | 113,011 | Anthropic | | 58 | 70 | Amazon Nova Micro 1.0 | 1197 | +5/-5 | 11,643 | Amazon | | 61 | 70 | Hunyuan-Standard-256K | 1189 | +10/-11 | 2,900 | Tencent | | 62 | 58 | Gemma-2-9B-it | 1191 | +4/-3 | 50,951 | Google | | 62 | 56 | Command R+ (04-2024) | 1190 | +3/-3 | 80,852 | Cohere | | 62 | 71 | Llama-3.1-Tulu-3-8B | 1185 | +9/-8 | 3,079 | Ai2 | | 63 | 57 | Qwen2-72B-Instruct | 1187 | +3/-3 | 38,880 | Alibaba | | 63 | 44 | GPT-4-0314 | 1186 | +4/-3 | 55,956 | OpenAI | | 63 | 70 | Ministral-8B-2410 | 1182 | +9/-6 | 5,114 | Mistral | | 65 | 61 | Command R (08-2024) | 1180 | +5/-5 | 10,846 | Cohere | | 65 | 71 | Aya-Expanse-8B | 1179 | +7/-8 | 9,571 | Cohere | | 67 | 62 | Claude 3 Haiku | 1179 | +2/-2 | 122,299 | Anthropic | | 67 | 56 | DeepSeek-Coder-V2-Instruct | 1178 | +4/-5 | 15,754 | DeepSeek AI | | 67 | 70 | Jamba-1.5-Mini | 1176 | +6/-6 | 9,270 | AI21 Labs | | 67 | 85 | Meta-Llama-3.1-8B-Instruct | 1176 | +3/-2 | 52,651 | Meta | | 76 | 56 | GPT-4-0613 | 1163 | +3/-3 | 91,646 | OpenAI | | 76 | 70 | Qwen1.5-110B-Chat | 1161 | +4/-3 | 27,466 | Alibaba | | 76 | 85 | Yi-1.5-34B-Chat | 1157 | +4/-4 | 25,126 | 01 AI | | 76 | 103 | QwQ-32B-Preview | 1153 | +11/-11 | 3,412 | Alibaba | | 77 | 70 | Mistral-Large-2402 | 1157 | +2/-3 | 64,925 | Mistral | | 77 | 71 | Reka-Flash-21B-online | 1156 | +4/-4 | 16,034 | Reka AI | | 78 | 93 | InternLM2.5-20B-chat | 1149 | +6/-6 | 10,595 | InternLM | | 79 | 78 | Llama-3-8B-Instruct | 1152 | +2/-2 | 109,237 | Meta | | 80 | 79 | Granite-3.1-8B-Instruct | 1142 | +10/-12 | 3,064 | IBM | | 81 | 75 | Command R (04-2024) | 1149 | +3/-3 | 56,375 | Cohere | | 81 | 77 | Mistral Medium | 1148 | +4/-3 | 35,554 | Mistral | | 81 | 76 | Reka-Flash-21B | 1148 | +3/-4 | 25,798 | Reka AI | | 81 | 73 | Mixtral-8x22b-Instruct-v0.1 | 1148 | +3/-3 | 53,787 | Mistral | | 81 | 75 | Qwen1.5-72B-Chat | 1147 | +3/-3 | 40,630 | Alibaba | | 83 | 93 | Gemma-2-2b-it | 1143 | +3/-3 | 42,631 | Google | | 90 | 73 | Gemini-1.0-Pro-001 | 1131 | +4/-4 | 18,792 | Google | | 90 | 83 | Zephyr-ORPO-141b-A35b-v0.1 | 1127 | +6/-7 | 4,863 | HuggingFace | | 91 | 87 | Qwen1.5-32B-Chat | 1125 | +5/-4 | 22,759 | Alibaba | | 91 | 96 | Granite-3.1-2B-Instruct | 1118 | +10/-10 | 3,127 | IBM | | 92 | 93 | Phi-3-Medium-4k-Instruct | 1123 | +3/-4 | 26,110 | Microsoft | | 92 | 105 | Starling-LM-7B-beta | 1119 | +4/-5 | 16,672 | Nexusflow | | 95 | 94 | Mixtral-8x7B-Instruct-v0.1 | 1114 | +0/-0 | 76,133 | Mistral | | 95 | 98 | Yi-34B-Chat | 1111 | +5/-4 | 15,920 | 01 AI | | 95 | 85 | Gemini Pro | 1110 | +7/-8 | 6,559 | Google | | 97 | 96 | Qwen1.5-14B-Chat | 1109 | +5/-4 | 18,675 | Alibaba | | 97 | 97 | WizardLM-70B-v1.0 | 1106 | +6/-6 | 8,379 | Microsoft | | 97 | 83 | GPT-3.5-Turbo-0125 | 1106 | +3/-2 | 68,861 | OpenAI | | 97 | 102 | Meta-Llama-3.2-3B-Instruct | 1103 | +6/-7 | 8,403 | Meta | | 99 | 93 | DBRX-Instruct-Preview | 1103 | +4/-3 | 33,730 | Databricks | | 99 | 100 | Phi-3-Small-8k-Instruct | 1102 | +4/-3 | 18,478 | Microsoft | | 99 | 101 | Tulu-2-DPO-70B | 1099 | +7/-7 | 6,664 | AllenAI/UW | | 103 | 93 | Granite-3.0-8B-Instruct | 1093 | +6/-7 | 7,002 | IBM | | 105 | 98 | OpenChat-3.5-0106 | 1092 | +6/-6 | 12,984 | OpenChat | | 106 | 112 | Llama-2-70B-chat | 1093 | +2/-2 | 39,634 | Meta | | 106 | 105 | Vicuna-33B | 1091 | +4/-4 | 22,950 | LMSYS | | 106 | 98 | Snowflake Arctic Instruct | 1090 | +3/-3 | 34,173 | Snowflake | | 106 | 108 | Starling-LM-7B-alpha | 1088 | +6/-5 | 10,417 | UC Berkeley | | 106 | 113 | Nous-Hermes-2-Mixtral-8x7B-DPO | 1084 | +9/-8 | 3,834 | NousResearch | | 108 | 99 | Gemma-1.1-7B-it | 1084 | +4/-4 | 25,059 | Google | | 108 | 113 | NV-Llama2-70B-SteerLM-Chat | 1081 | +9/-12 | 3,637 | Nvidia | | 111 | 102 | DeepSeek-LLM-67B-Chat | 1077 | +9/-7 | 4,988 | DeepSeek AI | | 113 | 100 | OpenChat-3.5 | 1076 | +7/-8 | 8,111 | OpenChat | | 113 | 104 | OpenHermes-2.5-Mistral-7B | 1074 | +9/-7 | 5,091 | NousResearch | | 113 | 108 | Granite-3.0-2B-Instruct | 1074 | +7/-7 | 7,195 | IBM | | 114 | 119 | Mistral-7B-Instruct-v0.2 | 1072 | +5/-4 | 20,051 | Mistral | | 114 | 119 | Qwen1.5-7B-Chat | 1070 | +8/-7 | 4,869 | Alibaba | | 115 | 119 | Phi-3-Mini-4K-Instruct-June-24 | 1071 | +6/-4 | 12,818 | Microsoft | | 115 | 94 | GPT-3.5-Turbo-1106 | 1068 | +5/-4 | 17,033 | OpenAI | | 115 | 123 | Phi-3-Mini-4k-Instruct | 1066 | +4/-5 | 21,095 | Microsoft | | 115 | 115 | Dolphin-2.2.1-Mistral-7B | 1062 | +13/-13 | 1,713 | Cognitive Computations | | 115 | 118 | SOLAR-10.7B-Instruct-v1.0 | 1062 | +8/-9 | 4,287 | Upstage AI | | 119 | 124 | Llama-2-13b-chat | 1063 | +4/-4 | 19,736 | Meta | | 122 | 119 | WizardLM-13b-v1.2 | 1059 | +7/-7 | 7,177 | Microsoft | | 125 | 127 | Meta-Llama-3.2-1B-Instruct | 1054 | +7/-6 | 8,531 | Meta | | 126 | 125 | Zephyr-7B-beta | 1053 | +6/-6 | 11,332 | HuggingFace | | 126 | 123 | SmolLM2-1.7B-Instruct | 1047 | +12/-14 | 2,370 | HuggingFace | | 126 | 119 | MPT-30B-chat | 1046 | +11/-12 | 2,648 | MosaicML | | 126 | 124 | Zephyr-7B-alpha | 1042 | +13/-16 | 1,815 | HuggingFace | | 126 | 125 | CodeLlama-70B-instruct | 1041 | +14/-20 | 1,193 | Meta | | 128 | 126 | CodeLlama-34B-instruct | 1043 | +7/-8 | 7,512 | Meta | | 128 | 119 | falcon-180b-chat | 1034 | +16/-16 | 1,326 | TII | | 131 | 120 | Vicuna-13B | 1042 | +4/-4 | 19,786 | LMSYS | | 131 | 126 | Gemma-7B-it | 1037 | +7/-5 | 9,174 | Google | | 131 | 127 | Phi-3-Mini-128k-Instruct | 1037 | +5/-4 | 21,627 | Microsoft | | 131 | 120 | Qwen-14B-Chat | 1035 | +8/-7 | 5,072 | Alibaba | | 131 | 143 | Llama-2-7B-chat | 1037 | +4/-5 | 14,553 | Meta | | 131 | 129 | Guanaco-33B | 1033 | +10/-8 | 3,000 | UW | | 140 | 132 | Gemma-1.1-2b-it | 1021 | +5/-7 | 11,346 | Google | | 140 | 136 | StripedHyena-Nous-7B | 1017 | +7/-8 | 5,271 | Together AI | | 141 | 150 | OLMo-7B-instruct | 1015 | +7/-7 | 6,508 | Allen AI | | 143 | 141 | Mistral-7B-Instruct-v0.1 | 1008 | +5/-6 | 9,144 | Mistral | | 144 | 143 | Vicuna-7B | 1005 | +6/-8 | 7,016 | LMSYS | | 144 | 132 | PaLM-Chat-Bison-001 | 1004 | +6/-5 | 8,743 | Google | | 148 | 147 | Gemma-2B-it | 989 | +9/-9 | 4,921 | Google | | 149 | 147 | Qwen1.5-4B-Chat | 988 | +5/-6 | 7,814 | Alibaba | | 151 | 151 | Koala-13B | 964 | +6/-7 | 7,036 | UC Berkeley | | 151 | 151 | ChatGLM3-6B | 955 | +9/-7 | 4,764 | Tsinghua | | 153 | 151 | GPT4All-13B-Snoozy | 932 | +11/-14 | 1,786 | Nomic AI | | 153 | 151 | MPT-7B-Chat | 928 | +10/-11 | 4,013 | MosaicML | | 153 | 156 | ChatGLM2-6B | 924 | +11/-10 | 2,707 | Tsinghua | | 153 | 156 | RWKV-4-Raven-14B | 922 | +8/-10 | 4,934 | RWKV | | 157 | 151 | Alpaca-13B | 902 | +7/-9 | 5,876 | Stanford | | 157 | 157 | OpenAssistant-Pythia-12B | 894 | +6/-8 | 6,381 | OpenAssistant | | 158 | 159 | ChatGLM-6B | 879 | +8/-11 | 4,988 | Tsinghua | | 159 | 159 | FastChat-T5-3B | 868 | +7/-11 | 4,299 | LMSYS | | 161 | 162 | StableLM-Tuned-Alpha-7B | 841 | +9/-12 | 3,341 | Stability AI | | 161 | 159 | Dolly-V2-12B | 822 | +8/-11 | 3,485 | Databricks | | 162 | 160 | LLaMA-13B | 800 | +14/-14 | 2,444 | Meta | ## 说明 * 排名(UB)：基于 Bradley-Terry 模型的上界排名 * 排名(StyleCtrl)：考虑对话风格的样式控制排名 * 置信区间：模型表现的置信区间 * 分数：基于模型性能的竞技场得分 ## 数据来源 * 数据来自 [lmarena.ai](https://lmarena.ai/)