자동화된 페이지입니다. AI가 자동으로 작성하고 번역했으며 소프트웨어가 출처와 대조해 검증했습니다. 사람 편집자는 없습니다. SI-IQ는 공개 벤치마크 결과로 만든 지수이며 IQ 검사가 아닙니다. 작동 방식
SimpleQA Verified: AI 모델 결과
SimpleQA Verified은(는) Epoch AI에서 실행하는 지식 벤치마크입니다. 당사 데이터에서 81개 모델이 발표된 결과를 가지고 있습니다.
SI-IQ 지수에 포함됩니다(적재량 k = 0.63, 우연 수준 0%).
- 평가 기관
- Epoch AI
- 벤치마크 홈페이지
- epoch.ai/benchmarks/simple-qa-verified
결과
| 1 | Gemini 4 Argon Google DeepMind | 76.0% | 140137–143 | 8 |
| 2 | GPT-6 Astra OpenAI | 75.6% | 150146–153 | 8 |
| 3 | GPT-6.1 Sol OpenAI | 73.9% | 148144–152 | 8 |
| 4 | Gemini 3.1 Pro Google DeepMind | 73.5% | 127124–130 | 9 |
| 5 | Claude Opus 5.5 Anthropic | 72.2% | 144141–148 | 8 |
| 6 | Claude Fable 5.1 Anthropic | 70.8% | 140136–143 | 8 |
| 7 | Claude Fable 5 Anthropic | 70.7% | 138134–141 | 8 |
| 8 | GPT-5.6 Sol OpenAI | 69.7% | 140137–143 | 8 |
| 9 | Gemini 3.8 Flash Google DeepMind | 69.7% | 130127–133 | 8 |
| 10 | Gemini 3.7 Flash Google DeepMind | 69.2% | 131128–134 | 8 |
| 11 | Gemini 3 Flash Google DeepMind | 66.8% | 122119–125 | 9 |
| 12 | Gemini 3.5 Flash Google DeepMind | 66.2% | 126123–129 | 9 |
| 13 | Gemini 3.6 Flash Google DeepMind | 66.2% | 125122–128 | 8 |
| 14 | GPT-5.5 OpenAI | 63.0% | 138135–141 | 9 |
| 15 | GPT-6 Sol OpenAI | 60.7% | 140137–144 | 8 |
| 16 | Muse Spark 1.2 Meta AI | 60.3% | 121112–129 | 2 데이터 부족 |
| 17 | Claude Opus 5 Anthropic | 59.9% | 136133–140 | 8 |
| 18 | Muse Spark 1.1 Meta AI | 57.8% | 125116–133 | 2 데이터 부족 |
| 19 | Muse Spark Open오픈 Meta AI | 56.6% | 127124–130 | 6 |
| 20 | Grok 4.7 xAI | 56.0% | 123120–127 | 7 데이터 부족 |
| 21 | Qwen3.7-Max Alibaba | 55.8% | 125122–128 | 9 |
| 22 | Claude Opus 4.8 Anthropic | 53.0% | 132129–135 | 8 |
| 23 | DeepSeek V4 Pro 0813오픈 DeepSeek | 52.9% | 128125–131 | 8 |
| 24 | Qwen 3.6 Max (Preview) Alibaba | 52.0% | 118112–124 | 5 데이터 부족 |
| 25 | Claude Opus 4.7 Anthropic | 51.7% | 127124–130 | 9 |
| 26 | Kimi K3오픈 Moonshot | 50.6% | 127124–130 | 8 |
| 27 | GPT-5 OpenAI | 50.1% | 121118–124 | 11 |
| 28 | o3 OpenAI | 49.4% | 116113–119 | 10 |
| 29 | Grok 4.6 xAI | 49.3% | 126123–129 | 8 |
| 30 | Qwen3-Max Alibaba | 48.7% | 106103–110 | 6 데이터 부족 |
| 31 | Grok 4.5 xAI | 48.3% | 124121–127 | 6 데이터 부족 |
| 32 | GPT-5.1 OpenAI | 48.0% | 120115–125 | 7 |
| 33 | Qwen3.8 Max (0902) Alibaba | 47.3% | 127123–130 | 6 데이터 부족 |
| 34 | Claude Opus 4.6 Anthropic | 47.0% | 123120–126 | 9 |
| 35 | DeepSeek-V4-Pro오픈 DeepSeek | 47.0% | 117113–120 | 9 |
| 36 | Claude Sonnet 5.5 Anthropic | 46.5% | 140137–143 | 8 |
| 37 | GPT-5.4 Pro OpenAI | 46.3% | 135131–138 | 5 데이터 부족 |
| 38 | Qwen 3.8 Max Alibaba | 45.8% | 130127–133 | 7 데이터 부족 |
| 39 | Claude Opus 4.5 Anthropic | 45.7% | 115111–118 | 10 |
| 40 | GPT-5.4 OpenAI | 45.1% | 131128–133 | 9 |
| 41 | Qwen 3.6 Plus Alibaba | 44.1% | 113109–117 | 7 데이터 부족 |
| 42 | GPT-5.6 Terra OpenAI | 43.2% | 134131–137 | 8 |
| 43 | DeepSeek V4.1 Flash오픈 DeepSeek | 42.9% | 128125–131 | 8 |
| 44 | GPT-6 Luna OpenAI | 41.4% | 130127–133 | 8 |
| 45 | o1 OpenAI | 41.1% | 10399–106 | 6 데이터 부족 |
| 46 | GLM-5.3오픈 Z.ai (Zhipu AI) | 41.0% | 125122–128 | 8 |
| 47 | GPT-5.6 Luna OpenAI | 41.0% | 130127–133 | 8 |
| 48 | Qwen3-235B-A22B-Thinking (Jul 2025)오픈 Alibaba | 40.4% | 111105–118 | 5 |
| 49 | Inkling오픈 Thinking Machines | 40.3% | 114110–117 | 7 |
| 50 | GPT-5.2 OpenAI | 37.1% | 125122–128 | 10 |
| 51 | Kimi K2.7 Code오픈 Moonshot | 36.5% | 118115–122 | 7 |
| 52 | Claude Sonnet 4.6 Anthropic | 35.5% | 113108–118 | 7 |
| 53 | Kimi K2.6오픈 Moonshot | 34.9% | 121118–124 | 9 |
| 54 | Kimi K2.5오픈 Moonshot | 34.3% | 119112–125 | 5 |
| 55 | GLM-5.2오픈 Z.ai (Zhipu AI) | 34.2% | 121118–124 | 9 |
| 56 | GLM-5.1오픈 Z.ai (Zhipu AI) | 34.0% | 116112–120 | 7 |
| 57 | Claude Sonnet 5 Anthropic | 33.7% | 126123–129 | 8 |
| 58 | DeepSeek V4 Flash 0731오픈 DeepSeek | 33.6% | 124121–127 | 8 |
| 59 | Grok 4.3 Beta xAI | 33.2% | 117114–121 | 6 데이터 부족 |
| 60 | GLM-4.7오픈 Z.ai (Zhipu AI) | 32.2% | 112107–118 | 5 |
| 61 | GPT-4.1 OpenAI | 31.1% | 9187–94 | 10 |
| 62 | Claude Sonnet 4.5 Anthropic | 30.7% | 110107–113 | 11 |
| 63 | GPT-5.4 Mini OpenAI | 29.4% | 117113–120 | 8 |
| 64 | GPT-4o (Aug 2024) OpenAI | 26.0% | 7368–77 | 8 |
| 65 | Qwen 3.5 Plus (hosted 397B-A17B) Alibaba | 25.4% | 109103–115 | 5 데이터 부족 |
| 66 | Claude Haiku 5.5 Anthropic | 23.8% | 129126–132 | 8 |
| 67 | GPT-5 mini OpenAI | 21.6% | 115112–118 | 11 |
| 68 | Qwen 3.5 Flash (hosted 35B-A3B) Alibaba | 20.3% | 107103–111 | 6 데이터 부족 |
| 69 | Mistral Large 4 Mistral AI | 20.0% | 122119–125 | 7 데이터 부족 |
| 70 | o4-mini OpenAI | 19.6% | 112109–115 | 10 |
| 71 | Inkling-Small오픈 Thinking Machines | 19.1% | 118115–122 | 7 |
| 72 | Qwen 3.6 Flash Alibaba | 15.9% | 107103–111 | 6 데이터 부족 |
| 73 | o3-mini OpenAI | 15.3% | 105101–108 | 9 |
| 74 | Claude Haiku 4.5 Anthropic | 13.2% | 10499–108 | 7 |
| 75 | GPT-4.1 mini OpenAI | 12.7% | 9188–95 | 10 |
| 76 | Claude 3 Opus Anthropic | 12.6% | 6560–70 | 5 데이터 부족 |
| 77 | GPT-5 nano OpenAI | 11.7% | 105102–109 | 10 |
| 78 | GPT-5.4 Nano OpenAI | 11.7% | 118115–121 | 8 |
| 79 | Gemma 4 31B IT오픈 Google DeepMind | 10.4% | 10498–110 | 4 데이터 부족 |
| 80 | GPT-4o mini OpenAI | 8.3% | 6864–73 | 9 |
| 81 | GPT-4.1 nano OpenAI | 6.0% | 7672–80 | 8 |
데이터 출처: Aider polyglot leaderboard data (Apache-2.0); Arena leaderboard dataset on Hugging Face (CC BY 4.0); Berkeley Function Calling Leaderboard, Gorilla project (Apache-2.0); Epoch AI, ‘Capabilities & benchmarking’ (CC BY 4.0); API list prices via LiteLLM's price table (MIT); each price links to the developer's pricing page; Source: OpenRouter (openrouter.ai), licensed under CC BY 4.0: reuse and republish with attribution to OpenRouter..