| 1 | — | ZGLM-4.7 (Reasoning)開源Z AI | 41.7 | — | — | — |
| 2 | ▲ 2 | DDeepSeek V3.2 (Reasoning)開源DeepSeek | 41.2 | — | — | — |
| 3 | ▼ 1 | KKimi K2 Thinking開源Kimi | 40.3 | — | — | — |
| 4 | ▲ 1 | MMiniMax-M2.1開源MiniMax | 39.4 | — | — | — |
| 5 | ▼ 2 | XMiMo-V2-Flash (Reasoning)開源Xiaomi | 39.1 | — | — | — |
| 6 | — | MMiniMax-M2開源MiniMax | 35.7 | — | — | — |
| 7 | ▲ 18 | ZGLM-4.7 (Non-reasoning)開源Z AI | 33.8 | — | — | — |
| 8 | ▲ 1 | DDeepSeek V3.1 Terminus (Reasoning)開源DeepSeek | 33.5 | — | — | — |
| 9 | ▼ 2 | Ogpt-oss-120B (high)開源OpenAI | 32.9 | — | — | — |
| 10 | ▲ 2 | DDeepSeek V3.2 Exp (Reasoning)開源DeepSeek | 32.5 | — | — | — |
| 11 | ▲ 2 | ZGLM-4.6 (Reasoning)開源Z AI | 32.2 | — | — | — |
| 12 | 新 | LK-EXAONE (Reasoning)開源LG AI Research | 31.9 | — | — | — |
| 13 | ▲ 4 | DDeepSeek V3.2 (Non-reasoning)開源DeepSeek | 31.9 | — | — | — |
| 14 | ▲ 20 | XMiMo-V2-Flash (Non-reasoning)開源Xiaomi | 30.1 | — | — | — |
| 15 | ▲ 27 | ZGLM-4.6 (Non-reasoning)開源Z AI | 29.8 | — | — | — |
| 16 | ▼ 6 | AQwen3 235B A22B 2507 (Reasoning)開源Alibaba | 29.4 | — | — | — |
| 17 | ▼ 9 | DDeepSeek V3.2 Speciale開源DeepSeek | 29.0 | — | — | — |
| 18 | ▲ 8 | KKimi K2 0905開源Kimi | 28.0 | — | — | — |
| 19 | ▲ 13 | DDeepSeek V3.2 Exp (Non-reasoning)開源DeepSeek | 28.0 | — | — | — |
| 20 | ▲ 15 | DDeepSeek V3.1 Terminus (Non-reasoning)開源DeepSeek | 28.0 | — | — | — |
| 21 | ▼ 5 | DDeepSeek V3.1 (Reasoning)開源DeepSeek | 27.9 | — | — | — |
| 22 | ▼ 11 | SApriel-v1.6-15B-Thinker開源ServiceNow | 27.8 | — | — | — |
| 23 | ▲ 18 | DDeepSeek V3.1 (Non-reasoning)開源DeepSeek | 27.7 | — | — | — |
| 24 | ▼ 10 | AQwen3 VL 235B A22B (Reasoning)開源Alibaba | 27.5 | — | — | — |
| 25 | ▼ 6 | DDeepSeek R1 0528 (May '25)開源DeepSeek | 27.1 | — | — | — |
| 26 | ▼ 2 | ZGLM-4.5 (Reasoning)開源Z AI | 26.8 | — | — | — |
| 27 | ▼ 12 | AQwen3 Next 80B A3B (Reasoning)開源Alibaba | 26.5 | — | — | — |
| 28 | ▼ 6 | BSeed-OSS-36B-Instruct開源ByteDance Seed | 25.0 | — | — | — |
| 29 | ▲ 21 | AQwen3 Coder 480B A35B Instruct開源Alibaba | 24.8 | — | — | — |
| 30 | ▼ 9 | NNVIDIA Nemotron 3 Nano 30B A3B (Reasoning)開源NVIDIA | 24.7 | — | — | — |
| 31 | ▼ 13 | Ogpt-oss-20B (high)開源OpenAI | 24.7 | — | — | — |
| 32 | ▲ 4 | AQwen3 235B A22B 2507 Instruct開源Alibaba | 24.6 | — | — | — |
| 33 | — | MMiniMax M1 80k開源MiniMax | 24.3 | — | — | — |
| 34 | ▲ 9 | NHyperCLOVA X SEED Think (32B)開源Naver | 24.0 | — | — | — |
| 35 | ▲ 21 | ZGLM-4.6V (Reasoning)開源Z AI | 23.8 | — | — | — |
| 36 | ▼ 7 | Ogpt-oss-120B (low)開源OpenAI | 23.8 | — | — | — |
| 37 | ▼ 10 | ZGLM-4.5-Air開源Z AI | 23.6 | — | — | — |
| 38 | 新 | KMi:dm K 2.5 Pro開源Korea Telecom | 23.2 | — | — | — |
| 39 | 新 | LK-EXAONE (Non-reasoning)開源LG AI Research | 23.2 | — | — | — |
| 40 | ▼ 10 | AQwen3 30B A3B 2507 (Reasoning)開源Alibaba | 22.8 | — | — | — |
| 41 | ▲ 13 | DDeepSeek V3 0324開源DeepSeek | 22.6 | — | — | — |
| 42 | ▲ 24 | MMistral Large 3開源Mistral | 22.5 | — | — | — |
| 43 | 新 | PINTELLECT-3開源Prime Intellect | 22.2 | — | — | — |
| 44 | ▲ 30 | MDevstral 2開源Mistral | 21.7 | — | — | — |
| 45 | ▼ 1 | Ogpt-oss-20B (low)開源OpenAI | 21.2 | — | — | — |
| 46 | ▼ 15 | MK2-V2 (high)開源MBZUAI Institute of Foundation Models | 21.1 | — | — | — |
| 47 | ▼ 2 | AQwen3 VL 235B A22B Instruct開源Alibaba | 20.8 | — | — | — |
| 48 | ▼ 9 | AQwen3 Next 80B A3B Instruct開源Alibaba | 20.3 | — | — | — |
| 49 | ▲ 31 | AQwen3 Coder 30B A3B Instruct開源Alibaba | 20.2 | — | — | — |
| 50 | ▲ 2 | AQwen3 235B A22B (Reasoning)開源Alibaba | 19.9 | — | — | — |
| 51 | ▼ 11 | ILing-1T開源InclusionAI | 19.7 | — | — | — |
| 51 | ▼ 14 | AQwen3 VL 30B A3B (Reasoning)開源Alibaba | 19.7 | — | — | — |
| 53 | ▼ 15 | NLlama Nemotron Super 49B v1.5 (Reasoning)開源NVIDIA | 19.4 | — | — | — |
| 54 | ▲ 5 | MK2-V2 (medium)開源MBZUAI Institute of Foundation Models | 19.2 | — | — | — |
| 55 | ▲ 33 | MDevstral Small 2開源Mistral | 19.1 | — | — | — |
| 56 | ▲ 22 | MLlama 4 Maverick開源Meta | 18.9 | — | — | — |
| 57 | ▼ 10 | AQwen3 4B 2507 (Reasoning)開源Alibaba | 18.8 | — | — | — |
| 58 | ▼ 12 | DDeepSeek R1 (Jan '25)開源DeepSeek | 18.7 | — | — | — |
| 59 | ▲ 46 | MLlama 3.1 Instruct 405B開源Meta | 17.5 | — | — | — |
| 60 | ▲ 32 | ZGLM-4.6V (Non-reasoning)開源Z AI | 17.4 | — | — | — |
| 61 | ▲ 35 | AQwen3 235B A22B (Non-reasoning)開源Alibaba | 17.3 | — | — | — |
| 62 | ▼ 7 | AQwen3 VL 32B Instruct開源Alibaba | 17.1 | — | — | — |
| 63 | — | AQwen3 32B (Reasoning)開源Alibaba | 17.1 | — | — | — |
| 64 | ▲ 21 | AQwen3 VL 8B (Reasoning)開源Alibaba | 17.0 | — | — | — |
| 65 | ▲ 19 | DDeepSeek V3 (Dec '24)開源DeepSeek | 17.0 | — | — | — |
| 66 | ▼ 17 | LEXAONE 4.0 32B (Reasoning)開源LG AI Research | 17.0 | — | — | — |
| 67 | ▲ 10 | AQwen3 14B (Reasoning)開源Alibaba | 16.8 | — | — | — |
| 68 | ▼ 3 | AQwen3 VL 30B A3B Instruct開源Alibaba | 16.3 | — | — | — |
| 69 | ▲ 22 | MMinistral 14B (Dec '25)開源Mistral | 16.3 | — | — | — |
| 70 | 新 | TFalcon-H1R-7B開源TII UAE | 16.2 | — | — | — |
| 71 | ▲ 2 | AQwen3 30B A3B (Reasoning)開源Alibaba | 15.8 | — | — | — |
| 72 | ▼ 12 | AQwen3 Omni 30B A3B (Reasoning)開源Alibaba | 15.8 | — | — | — |
| 73 | ▲ 37 | MDevstral Small (Jul '25)開源Mistral | 15.7 | — | — | — |
| 74 | ▼ 4 | NNVIDIA Nemotron Nano 9B V2 (Reasoning)開源NVIDIA | 15.5 | — | — | — |
| 75 | ▲ 40 | NLlama Nemotron Super 49B v1.5 (Non-reasoning)開源NVIDIA | 15.4 | — | — | — |
| 75 | ▼ 3 | AQwen3 30B A3B 2507 Instruct開源Alibaba | 15.4 | — | — | — |
| 77 | ▲ 23 | MMistral Small 3.2開源Mistral | 15.3 | — | — | — |
| 78 | ▲ 11 | MK2-V2 (low)開源MBZUAI Institute of Foundation Models | 15.3 | — | — | — |
| 79 | ▼ 22 | NNVIDIA Nemotron Nano 12B v2 VL (Reasoning)開源NVIDIA | 15.1 | — | — | — |
| 80 | ▲ 23 | MMinistral 8B (Dec '25)開源Mistral | 15.1 | — | — | — |
| 81 | ▲ 26 | MLlama 3.3 Instruct 70B開源Meta | 15.1 | — | — | — |
| 82 | ▲ 29 | AQwen3 VL 8B Instruct開源Alibaba | 14.6 | — | — | — |
| 83 | 新 | AOlmo 3.1 32B Think開源Allen Institute for AI | 14.2 | — | — | — |
| 84 | ▲ 20 | MLlama 4 Scout開源Meta | 14.0 | — | — | — |
| 85 | ▲ 47 | NLlama 3.1 Nemotron Instruct 70B開源NVIDIA | 14.0 | — | — | — |
| 86 | ▲ 38 | NNVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)開源NVIDIA | 13.9 | — | — | — |
| 87 | ▼ 11 | NNVIDIA Nemotron Nano 9B V2 (Non-reasoning)開源NVIDIA | 13.8 | — | — | — |
| 88 | ▲ 24 | CCommand A開源Cohere | 13.7 | — | — | — |
| 89 | ▲ 10 | AQwen3 14B (Non-reasoning)開源Alibaba | 13.4 | — | — | — |
| 90 | ▲ 30 | ZGLM-4.5V (Non-reasoning)開源Z AI | 13.2 | — | — | — |
| 91 | ▲ 2 | AQwen3 4B 2507 Instruct開源Alibaba | 13.1 | — | — | — |
| 92 | ▲ 24 | AQwen3 30B A3B (Non-reasoning)開源Alibaba | 12.4 | — | — | — |
| 93 | ▲ 67 | MLlama 3.1 Instruct 8B開源Meta | 12.2 | — | — | — |
| 94 | ▲ 49 | MMinistral 3B (Dec '25)開源Mistral | 12.1 | — | — | — |
| 95 | ▼ 1 | LEXAONE 4.0 32B (Non-reasoning)開源LG AI Research | 12.0 | — | — | — |
| 96 | ▲ 39 | IGranite 4.0 H Small開源IBM | 11.5 | — | — | — |
| 97 | ▲ 36 | AQwen3 8B (Non-reasoning)開源Alibaba | 10.8 | — | — | — |
| 98 | ▼ 3 | AQwen3 Omni 30B A3B Instruct開源Alibaba | 10.7 | — | — | — |
| 99 | ▲ 42 | GGemma 3 27B Instruct開源Google | 10.5 | — | — | — |
| 100 | ▲ 25 | NNVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)開源NVIDIA | 10.4 | — | — | — |
| 101 | ▲ 45 | AJamba 1.7 Large開源AI21 Labs | 9.7 | — | — | — |
| 102 | ▲ 48 | GGemma 3 12B Instruct開源Google | 9.4 | — | — | — |
| 103 | ▲ 11 | LExaone 4.0 1.2B (Reasoning)開源LG AI Research | 8.6 | — | — | — |
| 104 | ▲ 45 | LExaone 4.0 1.2B (Non-reasoning)開源LG AI Research | 8.6 | — | — | — |
| 105 | ▲ 71 | IGranite 4.0 H 1B開源IBM | 8.4 | — | — | — |
| 106 | 新 | AOlmo 3 7B Instruct開源Allen Institute for AI | 8.2 | — | — | — |
| 107 | ▲ 61 | AJamba 1.7 Mini開源AI21 Labs | 7.9 | — | — | — |
| 108 | ▲ 76 | LLFM2 2.6B開源Liquid AI | 7.9 | — | — | — |
| 109 | ▲ 68 | IGranite 4.0 1B開源IBM | 7.7 | — | — | — |
| 110 | ▲ 49 | LLFM2 8B A1B開源Liquid AI | 7.3 | — | — | — |
| 111 | ▲ 87 | IGranite 4.0 350M開源IBM | 7.2 | — | — | — |
| 112 | ▲ 79 | LLFM2 1.2B開源Liquid AI | 6.8 | — | — | — |
| 113 | ▲ 59 | AQwen3 1.7B (Non-reasoning)開源Alibaba | 6.8 | — | — | — |
| 114 | ▲ 55 | GGemma 3 4B Instruct開源Google | 6.6 | — | — | — |
| 115 | ▲ 50 | GGemma 3n E4B Instruct開源Google | 6.3 | — | — | — |
| 116 | ▲ 73 | AQwen3 0.6B (Non-reasoning)開源Alibaba | 6.1 | — | — | — |
| 117 | ▲ 80 | IGranite 4.0 H 350M開源IBM | 6.0 | — | — | — |
