Benchmark
GSM8k
math
reasoning
text
Grade School Math 8K, a dataset of 8.5K high-quality linguistically diverse grade school math word problems requiring multi-step reasoning and elementary arithmetic operations.
语言EN
满分1
参评模型46
模型排名
| 名次 | 模型 | 机构 | 分数 | 来源 |
|---|---|---|---|---|
| 1 | Kimi K2 Instruct | Moonshot AI | 97.3 | 来源 ↗ |
| 2 | o1 | OpenAI | 97.1 | 来源 ↗ |
| 3 | GPT-4.5 | OpenAI | 97.0 | 来源 ↗ |
| 4 | Llama 3.1 405B Instruct | Meta | 96.8 | 来源 ↗ |
| 5 | Claude 3.5 Sonnet | Anthropic | 96.4 | 来源 ↗ |
| 6 | Claude 3.5 Sonnet | Anthropic | 96.4 | 来源 ↗ |
| 7 | Qwen2.5 32B Instruct | Alibaba Cloud / Qwen Team | 95.9 | 来源 ↗ |
| 8 | Gemma 3 27B | 95.9 | 来源 ↗ | |
| 9 | Qwen2.5 72B Instruct | Alibaba Cloud / Qwen Team | 95.8 | 来源 ↗ |
| 10 | DeepSeek-V2.5 | DeepSeek | 95.1 | 来源 ↗ |
| 11 | Claude 3 Opus | Anthropic | 95.0 | 来源 ↗ |
| 12 | Qwen2.5 14B Instruct | Alibaba Cloud / Qwen Team | 94.8 | 来源 ↗ |
| 13 | Nova Pro | Amazon | 94.8 | 来源 ↗ |
| 14 | Nova Lite | Amazon | 94.5 | 来源 ↗ |
| 15 | Gemma 3 12B | 94.4 | 来源 ↗ | |
| 16 | Qwen3 235B A22B | Alibaba Cloud / Qwen Team | 94.4 | 来源 ↗ |
| 17 | Mistral Large 2 | Mistral AI | 93.0 | 来源 ↗ |
| 18 | Claude 3 Sonnet | Anthropic | 92.3 | 来源 ↗ |
| 19 | Nova Micro | Amazon | 92.3 | 来源 ↗ |
| 20 | Kimi K2 Base | Moonshot AI | 92.1 | 来源 ↗ |
| 21 | Qwen2.5 7B Instruct | Alibaba Cloud / Qwen Team | 91.6 | 来源 ↗ |
| 22 | Llama 3.1 Nemotron 70B Instruct | NVIDIA | 91.4 | 来源 ↗ |
| 23 | Qwen2 72B Instruct | Alibaba Cloud / Qwen Team | 91.1 | 来源 ↗ |
| 24 | Qwen2.5-Coder 32B Instruct | Alibaba Cloud / Qwen Team | 91.1 | 来源 ↗ |
| 25 | Gemini 1.5 Pro | 90.8 | 来源 ↗ | |
| 26 | Grok-1.5 | xAI | 90.0 | 来源 ↗ |
| 27 | Gemma 3 4B | 89.2 | 来源 ↗ | |
| 28 | Claude 3 Haiku | Anthropic | 88.9 | 来源 ↗ |
| 29 | Qwen2.5-Omni-7B | Alibaba Cloud / Qwen Team | 88.7 | 来源 ↗ |
| 30 | Phi-3.5-MoE-instruct | Microsoft | 88.7 | 来源 ↗ |
| 31 | Phi 4 Mini | Microsoft | 88.6 | 来源 ↗ |
| 32 | Jamba 1.5 Large | AI21 Labs | 87.0 | 来源 ↗ |
| 33 | Gemini 1.5 Flash | 86.2 | 来源 ↗ | |
| 34 | Phi-3.5-mini-instruct | Microsoft | 86.2 | 来源 ↗ |
| 35 | Qwen2.5-Coder 7B Instruct | Alibaba Cloud / Qwen Team | 83.9 | 来源 ↗ |
| 36 | Qwen2 7B Instruct | Alibaba Cloud / Qwen Team | 82.3 | 来源 ↗ |
| 37 | Granite 3.3 8B Instruct | IBM | 80.9 | 来源 ↗ |
| 38 | Mistral Small 3 24B Base | Mistral AI | 80.7 | 来源 ↗ |
| 39 | Llama 3.2 3B Instruct | Meta | 77.7 | 来源 ↗ |
| 40 | Jamba 1.5 Mini | AI21 Labs | 75.8 | 来源 ↗ |
| 41 | Gemma 2 27B | 74.0 | 来源 ↗ | |
| 42 | Command R+ | Cohere | 70.7 | 来源 ↗ |
| 43 | IBM Granite 4.0 Tiny Preview | IBM | 70.1 | 来源 ↗ |
| 44 | Gemma 2 9B | 68.6 | 来源 ↗ | |
| 45 | Gemma 3 1B | 62.8 | 来源 ↗ | |
| 46 | Granite 3.3 8B Base | IBM | 59.0 | 来源 ↗ |