MATH-500
在评估大型语言模型（LLM）的数学推理能力时，MATH和MATH-500是两个备受关注的基准测试。尽管它们都旨在衡量模型的数学解题能力，但在发布者、发布目的、评测目标和对比结果等方面存在显著差异。
更新于 2026-03-19
1,728 次浏览
问题数量
500
发布机构
OpenAI
评测类别
数学推理
评测指标
Accuracy
支持语言
英文
难度等级
高难度
简介
OpenAI从MATH评测数据集中精选的500个更具代表性的数学评测基准
相关资源
MATH-500 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 MATH-500 model rankings and full benchmark leaderboard
Browse the latest scores, model modes, release dates, and parameter sizes for MATH-500.
MATH-500详细排名数据表格
| 排名 | 模型 | |||
|---|---|---|---|---|
| 1 | 98.8 | 2025-05-06 | 未知 | |
| 2 | Gemini 2.5-Pro Normal | 98.8 | 2025-06-05 | 未知 |
| 3 | Claude Opus 4 Normal | 98.2 | 2025-05-23 | 未知 |
| 4 | GLM-4.5 Thinking Level · Medium | 98.2 | 2025-07-28 | 3550 |
| 5 | OpenAI o3 Normal | 98.1 | 2025-04-16 | 未知 |
| 6 | GLM-4.5-Air Thinking Level · Medium | 98.1 | 2025-07-28 | 1060 |
| 7 | Qwen3-235B-A22B Thinking Level · Medium | 98 | 2025-04-28 | 2350 |
| 8 | DeepSeek-R1-0528 Thinking Level · Medium | 98 | 2025-05-28 | 6710 |
| 9 | OpenAI o3-mini (high) Normal | 97.9 | 2025-01-31 | 未知 |
| 10 | Claude Opus 4.6 Deep Thinking | 97.6 | 2026-02-05 | 未知 |
| 11 | Qwen3-8B Thinking Level · Medium | 97.4 | 2025-04-28 | 80 |
| 12 | Kimi K2 Normal | 97.4 | 2025-07-11 | 10000 |
| 13 | DeepSeek-R1 Normal | 97.3 | 2025-01-20 | 6710 |
| 14 | Qwen3-32B Thinking Level · Medium | 97.2 | 2025-04-28 | 320 |
| 15 | MiniMax-M1-80k Normal | 96.8 | 2025-06-16 | 4560 |
| 16 | Pangu Pro MoE Normal | 96.8 | 2025-06-30 | 719 |
| 17 | OpenAI o1 Normal | 96.4 | 2024-12-05 | 未知 |
| 18 | ERNIE-4.5-300B-A47B Normal | 96.4 | 2025-06-30 | 3000 |
| 19 | Kimi k1.5 (Long-CoT) Normal | 96.2 | 2025-01-22 | 未知 |
| 20 | 96.2 | 2025-02-25 | 未知 | |
| 21 | Hunyuan-T1 Normal | 96.2 | 2025-03-21 | 未知 |
| 22 | Qwen3-235B-A22B Normal | 96.2 | 2025-04-28 | 2350 |
| 23 | MiniMax-M1-40k Normal | 96 | 2025-06-16 | 4560 |
| 24 | OpenAI o3-mini Thinking Level · Medium | 95.8 | 2025-01-31 | 未知 |
| 25 | 95 | 2025-04-05 | 20000 | |
| 26 | Kimi k1.5 (Short-CoT) Normal | 94.6 | 2025-01-22 | 未知 |
| 27 | 94.5 | 2025-01-20 | 700 | |
| 28 | DeepSeek-V3-0324 Normal | 94 | 2025-03-24 | 6710 |
| 29 | Hunyuan-7B Normal | 93.7 | 2025-08-04 | 70 |
| 30 | GPT-4.1 Normal | 92.8 | 2025-04-14 | 未知 |
滚动或悬停加载剩余 13 条