순위표
지식 랭킹
사실 지식과 환각 저항: SimpleQA, MMLU-Pro, MMMLU.
종합767코딩469수학90추론483에이전트89지식482멀티모달452코딩 에이전트68과학 추론96다국어12글쓰기95채팅 어시스턴트95리서치99도구 호출45긴 컨텍스트116속도472가성비471오픈 웨이트183
482개 모델
| # | 모델명 | Podium Score | 속도 |
|---|---|---|---|
| 01 | Claude Mythos Preview Anthropic추론프리뷰 | 100.0 | 80 t/s |
| 02 | Gemini 3.1 Pro Google | 97.6 | 136 t/s |
| 03 | Gemini 3 Pro Google | 92.9 | 120 t/s |
| 4 | Gemini 4 Argon Google추론프리뷰 | 91.0 | 88 t/s |
| 5 | Mistral Large 4 Mistral추론 | 91.0 | 92 t/s |
| 6 | GPT-6.1 Sol OpenAI추론 | 89.0 | 75 t/s |
| 7 | Claude 5.5 Sonnet Anthropic추론 | 88.0 | 94 t/s |
| 8 | Gemini 3.1 Pro Preview Google추론프리뷰 | 88.0 | 105 t/s |
| 9 | NVIDIA Nemotron 3.5 Super-VL NVIDIA오픈소스 | 85.0 | 110 t/s |
| 10 | MiMo v2.6 Flash Xiaomi오픈소스 | 84.0 | 118 t/s |
| 11 | Claude 5.5 Haiku Anthropic추론 | 81.0 | 145 t/s |
| 12 | Gemini 3.1 Flash Lite Preview Google프리뷰 | 81.0 | 210 t/s |
| 13 | Claude Mythos 5 Anthropic추론프리뷰 | 75.0 | 50 t/s |
| 14 | GPT-Image 2 OpenAI | 70.0 | 25 t/s |
| 15 | Qwen3.7 Max Alibaba | 69.3 | 203 t/s |
| 16 | FLUX.2 Pro Black Forest Labs | 68.0 | 30 t/s |
| 17 | Claude Sonnet 4.6 (max) Anthropic | 67.0 | 50 t/s |
| 18 | FLUX.2 Max Independent | 65.0 | 35 t/s |
| 19 | Qwen3.8 Flash Next Alibaba오픈소스 | 64.0 | 168 t/s |
| 20 | 62.0 | 34 t/s | |
| 21 | GLM-5.3 Flash Z.ai오픈소스 | 60.5 | 182 t/s |
| 22 | FLUX.2 Dev Black Forest Labs오픈소스 | 60.0 | 28 t/s |
| 23 | Kling 3.0 Pro Kwai | 60.0 | 25 t/s |
| 24 | Claude Fable 5.1 Anthropic추론 | 58.0 | 85 t/s |
| 25 | Gemini 3.8 Flash Google추론 | 58.0 | 302 t/s |
| 26 | 58.0 | 315 t/s | |
| 27 | Gemini 3.8 Flash (Low Reasoning) Google추론 | 58.0 | 340 t/s |
| 28 | Wan 2.2 A14B Alibaba오픈소스 | 55.0 | 15 t/s |
| 29 | Grok 4.7 (high) xAI추론 | 55.0 | 80 t/s |
| 30 | Grok 4.7 xAI추론 | 55.0 | 95 t/s |
| 31 | Granite 4.2 30B Instruct IBM오픈소스 | 52.0 | 78 t/s |
| 32 | ElevenLabs Scribe v2 ElevenLabs | 50.0 | 240 t/s |
| 33 | LTX-2.3 Pro Lightricks | 50.0 | 18 t/s |
| 34 | FLUX.2 Klein 9B Black Forest Labs오픈소스 | 50.0 | 55 t/s |
| 35 | Simba 3.2 Independent추론 | 50.0 | 62 t/s |
| 36 | ElevenLabs v4 Turbo Independent | 50.0 | 220 t/s |
| 37 | GPT-5.5 Pro OpenAI | 50.0 | 60 t/s |
| 38 | Qwen3.6 Plus Alibaba | 48.4 | 55 t/s |
| 39 | Eleven Multilingual v3 ElevenLabs | 48.0 | 200 t/s |
| 40 | Agnes 2.5 Pro Beta Agnes | 48.0 | 65 t/s |
| 41 | Voxtral 3 TTS Mistral | 47.0 | 165 t/s |
| 42 | Step Audio 2 StepFun | 46.0 | 150 t/s |
| 43 | DeepSeek V4.1 Flash DeepSeek오픈소스추론 | 46.0 | 214 t/s |
| 44 | Kimi K3.1 Pro Moonshot AI오픈소스추론 | 45.0 | 45 t/s |
| 45 | Cartesia Ink Whisper Cartesia | 45.0 | 260 t/s |
| 46 | Qwen Audio 3.1 TTS+ Alibaba오픈소스 | 45.0 | 140 t/s |
| 47 | Claude Fable 5.1 (XHigh) Anthropic추론 | 45.0 | 78 t/s |
| 48 | GPT-6 Astra (xhigh) OpenAI추론 | 45.0 | 76 t/s |
| 49 | Claude Fable 5.1 (High) Anthropic추론 | 45.0 | 92 t/s |
| 50 | Deepgram Flux Independent | 45.0 | 260 t/s |
| 51 | LTX-2.3 Fast Lightricks | 45.0 | 60 t/s |
| 52 | GPT-6 Astra (high) OpenAI추론 | 45.0 | 90 t/s |
| 53 | Claude Opus 5.5 (XHigh) Anthropic추론 | 45.0 | 62 t/s |
| 54 | GPT-6 Luna (xhigh) OpenAI추론 | 45.0 | 118 t/s |
| 55 | Claude Opus 5.5 Anthropic추론 | 45.0 | 68 t/s |
| 56 | Claude Fable 5.1 (Medium) Anthropic추론 | 45.0 | 104 t/s |
| 57 | Claude Opus 5.5 (High) Anthropic추론 | 45.0 | 74 t/s |
| 58 | GPT-6 Luna (high) OpenAI추론 | 45.0 | 128 t/s |
| 59 | GPT-6 Astra (medium) OpenAI추론 | 45.0 | 98 t/s |
| 60 | Claude Fable 5.1 (Low) Anthropic추론 | 45.0 | 118 t/s |
| 61 | Claude Opus 5.5 (Medium) Anthropic추론 | 45.0 | 82 t/s |
| 62 | GPT-6 Luna (medium) OpenAI추론 | 45.0 | 138 t/s |
| 63 | GPT-6 Sol (xhigh) OpenAI추론 | 45.0 | 210 t/s |
| 64 | GPT-6 Sol OpenAI추론 | 45.0 | 220 t/s |
| 65 | GPT-6 Astra (low) OpenAI추론 | 45.0 | 110 t/s |
| 66 | GPT-6 Sol (high) OpenAI추론 | 45.0 | 226 t/s |
| 67 | Claude Opus 5.5 (Low) Anthropic추론 | 45.0 | 92 t/s |
| 68 | GPT-6 Luna (low) OpenAI추론 | 45.0 | 150 t/s |
| 69 | GPT-6 Sol (medium) OpenAI추론 | 45.0 | 238 t/s |
| 70 | 45.0 | 165 t/s | |
| 71 | GPT-6 Sol (low) OpenAI추론 | 45.0 | 250 t/s |
| 72 | 45.0 | 265 t/s | |
| 73 | Muse Spark 1.3 Muse추론 | 45.0 | 75 t/s |
| 74 | Step-5 StepFun추론 | 45.0 | 110 t/s |
| 75 | GLM-5.3 (low) Zhipu AI | 45.0 | 100 t/s |
| 76 | Mercury 2.5 Inception | 45.0 | 105 t/s |
| 77 | MiMo V2.6 Pro Xiaomi오픈소스 | 45.0 | 72 t/s |
| 78 | Gpt 5 6 Terra High OpenAI | 45.0 | 90 t/s |
| 79 | Granite 4.2 8B Instruct IBM오픈소스 | 45.0 | 115 t/s |
| 80 | Claude Sonnet 4 6 Adaptive Anthropic | 45.0 | 80 t/s |
| 81 | Gpt 5 6 Luna High OpenAI | 45.0 | 90 t/s |
| 82 | Gpt 5 6 Terra Medium OpenAI | 45.0 | 90 t/s |
| 83 | DeepSeek V4 Flash 0420 (Non-reasoning) DeepSeek오픈소스 | 45.0 | 210 t/s |
| 84 | Deepseek V4 Pro 0424 DeepSeek오픈소스 | 45.0 | 60 t/s |
| 85 | Claude Opus 4 6 Adaptive Anthropic | 45.0 | 80 t/s |
| 86 | Gpt 5 5 Low OpenAI | 45.0 | 90 t/s |
| 87 | Deepseek V4 Pro 0424 High DeepSeek오픈소스 | 45.0 | 60 t/s |
| 88 | Claude Opus 4 5 Thinking Anthropic추론 | 45.0 | 50 t/s |
| 89 | Gpt 5 6 Terra Low OpenAI | 45.0 | 90 t/s |
| 90 | Gpt 5 2 Codex OpenAI | 45.0 | 90 t/s |
| 91 | Qwen3 6 Max Alibaba오픈소스 | 45.0 | 85 t/s |
| 92 | 45.0 | 80 t/s | |
| 93 | Gpt 5 4 Low OpenAI | 45.0 | 90 t/s |
| 94 | Agnes 2 5 Pro Alpha Agnes | 45.0 | 60 t/s |
| 95 | Gpt 5 2 Medium OpenAI | 45.0 | 90 t/s |
| 96 | Gpt 5 6 Luna Medium OpenAI | 45.0 | 90 t/s |
| 97 | Deepseek V3 2 Reasoning DeepSeek오픈소스추론 | 45.0 | 40 t/s |
| 98 | Grok 4 20 xAI | 45.0 | 80 t/s |
| 99 | Claude 4 5 Sonnet Thinking Anthropic추론 | 45.0 | 50 t/s |
| 100 | 45.0 | 80 t/s |
Podium Score는 5개 리더보드의 가중 평균입니다. 자세한 내용은 평가 방법을 참고하세요.