加载中...
加载中...
一个包含 8500 道小学数学题的基准，用于评估模型的数学推理能力。
Source: DataLearnerAI
Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology
| 排名 | 模型 | |||
|---|---|---|---|---|
| 1 | ERNIE-4.5-300B-A47Bdefault | 96.6 | 2025-06-30 | 3000 |
| 2 | Qwen3-235B-A22Bdefault | 96.4 | 2025-04-28 | 2350 |
| 3 | DeepSeek-V3-0324default | 96.3 | 2025-03-24 | 6710 |
| 4 | Pangu Embeddeddefault | 95.98 | 2025-06-30 | 70 |
| 5 | Qwen2.5-32Bdefault | 95.9 | 2024-09-18 | 320 |
| 6 | GPT-4.1default | 95.9 | 2025-04-14 | 未知 |
| 7 | Claude3-Opusdefault | 95 | 2024-03-04 | 未知 |
| 8 | Qwen2.5-Maxdefault | 94.5 | 2025-01-28 | 未知 |
| 9 | Hunyuan-A13B-Instructdefault | 91.83 | 2025-06-27 | 800 |
| 10 | Qwen2.5-72Bdefault | 91.5 | 2024-09-18 | 727 |
| 11 | GPT-4o minidefault | 91.3 | 2024-07-18 | 未知 |
| 12 | Qwen3-Nextdefault | 90.3 | 2025-09-11 | 800 |
| 13 | Phi-4-mini-instruct (3.8B)default | 88.6 | 2025-02-27 | 38 |
| 14 | Qwen2.5-7Bdefault | 85.4 | 2024-09-18 | 70 |
| 15 | Llama3.1-8B-Instructdefault | 82.4 | 2024-07-23 | 80 |
| 16 | Qwen2.5-3Bdefault | 79.1 | 2024-09-18 | 30 |
| 17 | Moonlight-16B-A3B-Instructdefault | 77.4 | 2025-02-23 | 160 |
| 18 | Gemma 2 - 9Bdefault | 70.7 | 2024-06-27 | 90 |
| 19 | Llama3.1-8Bdefault | 55.3 | 2024-07-23 | 80 |
| 20 | Mistral-7B-Instruct-v0.3default | 36.2 | 2024-05-22 | 70 |
| 21 | Llama-3.2-3Bdefault | 34 | 2024-09-18 | 32 |
| 22 | Gemini 1.5 Prodefault | 0 | 2024-02-15 | 未知 |
| 23 | Llama3.1-405B Instructdefault | 0 | 2024-07-23 | 4050 |
| 24 | Amazon Nova Prodefault | 0 | 2024-12-03 | 未知 |