MMLU
更新于 2026-04-16
4,356 次浏览
问题数量
15000
发布机构
个人
评测类别
综合评估
评测指标
Accuracy
支持语言
英文
难度等级
基础
简介
一个涵盖 57 个主题的多项选择题基准，用于评估大规模语言模型的知识和推理能力。
相关资源
MMLU Model Score Leaderboard
Source: DataLearnerAI
Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology
模式说明:
normal
thinking
low
medium
high
deeper thinking
parallel_thinking
图表加载中...
Latest MMLU model rankings and full benchmark leaderboard
Browse the latest scores, model modes, release dates, and parameter sizes for MMLU.
Model release cutoff:
MMLU详细排名数据表格
| 排名 | 模型 | |||
|---|---|---|---|---|
| 1 | DeepSeek-V3.1 Thinking Enabled | 93.40 | 2025-08-20 | 6710 |
| 2 | OpenAI o4 - mini Thinking Enabled | 93.00 | 2025-04-16 | 未知 |
| 3 | Gemini 3.1 Pro Preview Thinking Level · High | 92.60 | 2026-02-20 | 未知 |
| 4 | OpenAI o1 Standard Mode | 91.80 | 2024-12-05 | 未知 |
| 5 | DeepSeek-V3.1 Standard Mode | 91.80 | 2025-08-20 | 6710 |
| 6 | Opus 4.7 Standard Mode | 91.50 | 2026-04-16 | 未知 |
| 7 | Claude Opus 4.6 Extended Thinking | 91.05 | 2026-02-05 | 未知 |
| 8 | DeepSeek-R1 Standard Mode | 90.80 | 2025-01-20 | 6710 |
| 9 | GPT-4.1 Standard Mode | 90.20 | 2025-04-14 | 未知 |
| 10 | GPT OSS 120B Thinking Enabled | 90.00 | 2025-08-06 | 117 |
| 11 | GPT-5.2 Thinking Level · Extra High | 89.60 | 2025-12-11 | 未知 |
| 12 | Hunyuan-TurboS Standard Mode | 89.50 | 2025-03-10 | 未知 |
| 13 | Kimi K2 Standard Mode | 89.50 | 2025-07-11 | 10000 |
| 14 | Pangu Pro MoE Standard Mode | 89.30 | 2025-06-30 | 719 |
| 15 | GPT-4o Standard Mode | 88.70 | 2024-05-13 | 未知 |
| 16 | Llama3.1-405B Instruct Standard Mode | 88.60 | 2024-07-23 | 4050 |
| 17 | DeepSeek-V3 Standard Mode | 88.50 | 2024-12-26 | 6810 |
| 18 | Claude 3.5 Sonnet Standard Mode | 88.30 | 2024-06-21 | 未知 |
| 19 | Claude 3.5 Sonnet New Standard Mode | 88.30 | 2024-10-22 | 未知 |
| 20 | Hunyuan-A13B-Instruct Standard Mode | 88.17 | 2025-06-27 | 800 |
| 21 | Qwen2.5-Max Standard Mode | 87.90 | 2025-01-28 | 未知 |
| 22 | Grok 2 Standard Mode | 87.50 | 2024-08-13 | 2690 |
| 23 | GPT-4.1 mini Standard Mode | 87.50 | 2025-04-14 | 未知 |
| 24 | Kimi k1.5 (Short-CoT) Standard Mode | 87.40 | 2025-01-22 | 未知 |
| 25 | Gemini 1.5 Pro Standard Mode | 87.10 | 2024-02-15 | 未知 |
| 26 | OpenAI o3-mini (high) Standard Mode | 86.90 | 2025-01-31 | 未知 |
| 27 | Claude3-Opus Standard Mode | 86.80 | 2024-03-04 | 未知 |
| 28 | Gemini 2.0 Pro Experimental Standard Mode | 86.50 | 2025-02-05 | 未知 |
| 29 | DeepSeek-V3-0324 Standard Mode | 86.50 | 2025-03-24 | 6710 |
| 30 | ERNIE-4.5-300B-A47B Standard Mode | 86.50 | 2025-06-30 | 3000 |
滚动或悬停加载剩余 35 条