🏆 Leaderboard
Model ranking from user votes (Bradley–Terry / Elo).
112 models · 163,474 votes · updated 2026-10-07 13:46:58 PDT
| Rank | Model | Arena Elo | Votes | Organization | License |
|---|---|---|---|---|---|
| 1 | Google: Gemini 3 Pro Preview | 1199 +26/−24 | 537 | proprietary | |
| 1 | Google: Gemini 3.5 Flash | 1191 +14/−13 | 2 138 | proprietary | |
| 1 | Google: Gemini Pro 2.5 Preview 03-25 | 1187 +23/−23 | 710 | proprietary | |
| 1 | Google: Gemini 3.1 Pro Preview | 1177 +11/−10 | 3 292 | proprietary | |
| 1 | Qwen3.6 Max Preview | 1164 +18/−17 | 1 119 | - | proprietary |
| 1 | GPT-5.4 | 1163 +16/−16 | 1 233 | proprietary | |
| 1 | Anthropic: Claude Opus 4.7 | 1154 +27/−29 | 423 | Anthropic | proprietary |
| 2 | OpenAI: GPT-5.5 | 1150 +27/−26 | 458 | proprietary | |
| 1 | Google: Gemini Flash 2.5 Preview 04-17:thinking | 1147 +34/−28 | 422 | proprietary | |
| 7 | GPT-5.2 | 1123 +14/−13 | 1 996 | OpenAI | proprietary |
| 7 | GigaChat 3.5 | 1121 +10/−9 | 4 318 | - | proprietary |
| 7 | DeepSeek V3.2 | 1119 +12/−12 | 1 971 | DeepSeek | open |
| 7 | GPT-5.1 | 1116 +15/−13 | 1 791 | OpenAI | proprietary |
| 7 | Anthropic: Claude Sonnet 4.6 | 1115 +25/−26 | 433 | Anthropic | proprietary |
| 7 | GPT-5 Chat | 1115 +14/−12 | 2 026 | OpenAI | proprietary |
| 7 | GPT-5.3 Chat | 1113 +19/−19 | 960 | - | proprietary |
| 7 | DeepSeek V3.1 | 1113 +17/−17 | 1 044 | DeepSeek | proprietary |
| 7 | Google: Gemini Flash 2.5 Preview 04-17 | 1112 +28/−28 | 441 | proprietary | |
| 10 | Qwen3 Max | 1104 +14/−14 | 1 809 | Qwen | proprietary |
| 10 | Claude Opus 4.1 | 1098 +21/−20 | 825 | Anthropic | proprietary |
| 10 | Claude Sonnet 4.5 | 1097 +19/−18 | 949 | Anthropic | proprietary |
| 12 | DeepSeek R1 0528 | 1093 +16/−15 | 1 337 | DeepSeek | proprietary |
| 15 | xAI: Grok 4 | 1084 +15/−15 | 1 435 | xAI | proprietary |
| 13 | TERA | 1084 +21/−22 | 646 | - | proprietary |
| 15 | gpt-4.1-2025-04-14 | 1083 +19/−17 | 933 | OpenAI | proprietary |
| 15 | DeepSeek R1 | 1076 +22/−20 | 789 | DeepSeek | open |
| 19 | DeepSeek: DeepSeek V3 0324 | 1076 +12/−13 | 2 028 | DeepSeek | proprietary |
| 18 | Xiaomi: MiMo-V2-Flash | 1071 +19/−17 | 863 | Xiaomi | open |
| 20 | Gemma 3 27B | 1070 +13/−12 | 1 825 | proprietary | |
| 20 | Alice AI LLM | 1069 +13/−12 | 1 929 | Yandex | proprietary |
| 19 | claude-sonnet-4-20250514 | 1066 +19/−20 | 949 | Anthropic | proprietary |
| 19 | Qwen3 235B A22B | 1066 +22/−21 | 658 | Qwen | proprietary |
| 20 | GPT-OSS-120B | 1064 +17/−15 | 1 183 | OpenAI | proprietary |
| 21 | claude-3-7-sonnet-20250219:thinking | 1063 +16/−16 | 1 138 | Anthropic | proprietary |
| 23 | gpt-4o-2024-11-20 | 1063 +13/−14 | 1 624 | OpenAI | proprietary |
| 20 | Google: Gemini Flash 2.0 | 1060 +20/−20 | 850 | proprietary | |
| 21 | DeepSeek V3 | 1059 +20/−20 | 869 | DeepSeek | open |
| 23 | o4-mini-2025-04-16 | 1052 +18/−16 | 930 | OpenAI | proprietary |
| 25 | claude-3-7-sonnet-20250219 | 1048 +16/−15 | 1 445 | Anthropic | proprietary |
| 27 | Google: Gemini Pro 1.5 | 1047 +15/−15 | 1 537 | proprietary | |
| 27 | Claude Haiku 4.5 | 1046 +13/−11 | 2 122 | Anthropic | proprietary |
| 27 | gpt-4.1-mini-2025-04-14 | 1040 +18/−18 | 926 | OpenAI | proprietary |
| 31 | Moonshot AI: Kimi K2 | 1039 +12/−12 | 2 271 | Moonshot AI | proprietary |
| 34 | Z.AI: GLM 4.7 | 1033 +13/−13 | 2 251 | Z.AI | proprietary |
| 36 | MiniMax: MiniMax M2.1 | 1031 +12/−13 | 2 093 | MiniMaxAI | open |
| 31 | Qwen3 14B | 1030 +19/−17 | 922 | Qwen | proprietary |
| 37 | claude-3-5-sonnet-20241022 | 1026 +13/−13 | 1 977 | Anthropic | proprietary |
| 36 | t-tech/T-pro-it-2.0 | 1024 +17/−18 | 1 040 | t-bank-ai | open |
| 42 | gpt-4o-2024-05-13 | 1019 +13/−13 | 2 111 | OpenAI | proprietary |
| 42 | o3-mini-2025-01-31 | 1018 +14/−12 | 1 757 | OpenAI | proprietary |
| 39 | o1-mini-2024-09-12 | 1017 +17/−17 | 1 096 | OpenAI | proprietary |
| 43 | Qwen3 8B | 1012 +11/−12 | 2 179 | Qwen | proprietary |
| 43 | claude-3-5-sonnet-20240620 | 1012 +14/−15 | 1 653 | Anthropic | proprietary |
| 42 | MiniMax: MiniMax-Text-01 | 1011 +20/−21 | 751 | MiniMaxAI | open |
| 44 | Qwen3 30B A3B | 1009 +12/−12 | 2 287 | Qwen | proprietary |
| 46 | Llama 4 Maverick | 1009 +10/−10 | 2 748 | Meta | open |
| 46 | GPT-OSS-20B | 1006 +13/−12 | 2 107 | OpenAI | proprietary |
| 46 | Qwen2.5 72B Instruct | 1000 +16/−16 | 1 247 | Qwen | open |
| 50 | Qwen3 32B | 993 +14/−14 | 1 730 | Qwen | proprietary |
| 48 | GigaChat 2 Max | 992 +18/−18 | 953 | Sber | proprietary |
| 51 | Llama 4 Scout | 991 +10/−10 | 2 714 | Meta | open |
| 51 | YandexGPT 5.1 Pro | 990 +13/−12 | 2 244 | Yandex | proprietary |
| 51 | gpt-4-turbo-2024-04-09 | 990 +12/−13 | 2 103 | OpenAI | proprietary |
| 56 | gpt-4.1-nano-2025-04-14 | 984 +11/−11 | 3 065 | OpenAI | proprietary |
| 58 | Llama 3.3 70B Instruct | 979 +9/−9 | 3 861 | Meta* | open |
| 57 | Google: Gemini Flash 1.5 | 979 +14/−16 | 1 366 | proprietary | |
| 55 | GigaChat 2 Pro | 978 +18/−18 | 929 | Sber | proprietary |
| 56 | RefalMachine/RuadaptQwen2.5-32B-Pro-Beta | 975 +20/−20 | 749 | RefalMachine | open |
| 59 | Llama 3.1 405B Instruct Turbo | 971 +11/−12 | 2 236 | Meta* | proprietary |
| 59 | gpt-4o-mini-2024-07-18 | 969 +13/−13 | 1 872 | OpenAI | proprietary |
| 59 | GigaChat-Max-preview 4.0.26.20 | 968 +14/−15 | 1 381 | Sber | proprietary |
| 62 | claude-3-5-haiku-20241022 | 966 +12/−14 | 1 986 | Anthropic | proprietary |
| 61 | Gemma 2 27B | 965 +13/−14 | 1 766 | proprietary | |
| 58 | t-tech/T-pro-it-1.0 | 963 +22/−20 | 547 | t-bank-ai | open |
| 61 | Vikhrmodels/Vikhr-Nemo-12B-Instruct-R-21-09-24 | 960 +18/−19 | 859 | Vikhrmodels | open |
| 67 | Cohere: Command R+ (08-2024) | 947 +16/−17 | 1 210 | Oracle | proprietary |
| 66 | Mistral Small 3 | 946 +18/−17 | 858 | Mistral | open |
| 72 | Qwen2.5 Coder 32B Instruct | 942 +10/−10 | 2 815 | Qwen | open |
| 71 | Llama 3.1 70B Instruct Turbo | 941 +12/−12 | 2 196 | Meta* | proprietary |
| 74 | YandexGPT Experimental | 935 +16/−16 | 1 185 | Yandex | proprietary |
| 71 | t-tech/T-lite-it-1.0 | 931 +23/−21 | 595 | t-bank-ai | open |
| 74 | saiga_llama3_70b | 931 +19/−23 | 676 | Ilya Gusev | open |
| 74 | GigaChat-Pro 4.0.26.20 | 930 +15/−17 | 1 082 | Sber | proprietary |
| 74 | GigaChat 2 Lite | 929 +19/−19 | 954 | Sber | proprietary |
| 75 | claude-3-haiku-20240307 | 929 +14/−13 | 1 594 | Anthropic | proprietary |
| 74 | T-lite-instruct-0.1 | 929 +15/−14 | 1 633 | t-bank-ai | open |
| 74 | YandexGPT 5 Pro | 926 +18/−19 | 921 | Yandex | proprietary |
| 76 | gpt-4-0613 | 922 +12/−13 | 2 079 | OpenAI | proprietary |
| 76 | LLaMA-3 Chat (70B) | 920 +12/−13 | 1 924 | Meta* | proprietary |
| 80 | Qwen 2 Instruct (72B) | 908 +13/−14 | 1 770 | Qwen | open |
| 83 | YandexGPT 4 Pro | 902 +12/−13 | 2 009 | Yandex | proprietary |
| 77 | GigaChat-Pro 4.0.26.15 | 901 +31/−30 | 319 | Sber | proprietary |
| 84 | GigaChat 4.0.26.20 | 896 +16/−15 | 1 335 | Sber | proprietary |
| 86 | GigaChat-Plus 4.0.26.15 | 892 +17/−17 | 1 153 | Sber | proprietary |
| 90 | saiga_llama3_8b_v7 | 890 +14/−14 | 1 550 | Ilya Gusev | open |
| 83 | GigaChat 4.0.26.15 | 885 +28/−28 | 369 | Sber | proprietary |
| 91 | Microsoft: Phi 4 | 880 +11/−12 | 2 830 | Microsoft | open |
| 91 | YandexGPT 3 Pro | 878 +13/−14 | 1 767 | Yandex | proprietary |
| 90 | saiga_phi3_medium | 876 +23/−26 | 497 | Ilya Gusev | open |
| 91 | RefalMachine/ruadapt_llama3_instruct_lep_saiga_kto_ablitirated | 874 +19/−21 | 758 | RefalMachine | open |
| 95 | LLaMA-3 Chat (8B) | 861 +12/−13 | 2 077 | Meta* | proprietary |
| 93 | Vikhrmodels/it-5.2-fp16-cp | 859 +20/−20 | 804 | Vikhrmodels | open |
| 95 | YandexGPT 3 Lite | 858 +13/−15 | 1 780 | Yandex | proprietary |
| 95 | Llama 3.2 11B Instruct | 858 +14/−16 | 1 302 | Meta* | open |
| 98 | gpt-3.5-turbo-0125 | 851 +12/−13 | 1 980 | OpenAI | proprietary |
| 98 | Llama 3.1 8B Instruct Turbo | 850 +12/−14 | 1 895 | Meta* | proprietary |
| 98 | MTSAIR/Cotype-Nano | 844 +18/−18 | 1 009 | MTSAIR | open |
| 102 | GigaChat-Pro 4.0.26.8 | 825 +19/−19 | 932 | Sber | proprietary |
| 105 | GigaChat-Pro 2.2.25.3 | 820 +19/−20 | 981 | Sber | proprietary |
| 107 | GigaChat 4.0.26.8 | 814 +18/−20 | 895 | Sber | proprietary |
| 109 | GigaChat 3.1.25.3 | 783 +18/−18 | 984 | Sber | proprietary |
| 111 | GigaChat-Plus 3.1.25.3 | 769 +18/−18 | 994 | Sber | proprietary |