YZ Index
评测数据
当前展示：Run #380 | 2026-10-15 | 128 题库 | 公式 v7 | 判分 v6.4
数据公开说明：为防止题库污染和过拟合，题目原文和预期答案不公开。本页展示模型回答、得分、判分方式等透明数据。完整方法论请参阅方法论页面。
| 模型 | 代码执行 | 材料约束 | 工程判断 | 任务表达 | 诚信 | 主榜分 | 性价比 | 稳定性 | 可用性 | 逐题 |
|---|---|---|---|---|---|---|---|---|---|---|
| Gemini 2.5 Pro gemini | 97.80 | 93.60 | 84.00 | 91.70 | 100.00 pass | 95.91 | 98.7 | 74.8 | 100.0 | |
| Claude Opus 4.7 claude | 97.00 | 89.30 | 88.90 | 45.60 | 100.00 pass | 93.54 | 28.9 | 47.8 | 100.0 | |
| Claude Sonnet 4.6 claude | 93.00 | 87.90 | 100.00 | 84.70 | 80.00 pass | 90.71 | 90.6 | 65.9 | 100.0 | |
| 豆包 Pro doubao | 90.70 | 87.90 | 100.00 | 91.70 | 75.00 pass | 89.44 | 100.0 | 51.8 | 100.0 | |
| GPT-5.5 gpt | 92.00 | 82.80 | 100.00 | 91.70 | 90.00 pass | 87.86 | 79.2 | 70.9 | 100.0 | |
| GPT-6 Sol gpt | 92.00 | 82.80 | 100.00 | 87.50 | 80.00 pass | 87.86 | 97.6 | 65.1 | 100.0 | |
| Grok 4 grok | 95.10 | 77.20 | 76.40 | 71.70 | 100.00 pass | 87.05 | 88.9 | 48.6 | 100.0 | |
| GPT-6 Luna gpt | 92.00 | 78.60 | 100.00 | 87.50 | 80.00 pass | 85.97 | 100.0 | 65.6 | 100.0 | |
| DeepSeek V4 Pro deepseek | 72.00 | 100.00 | 100.00 | 87.50 | 90.00 pass | 84.60 | 98.9 | 72.1 | 95.0 | |
| GPT-o3 gpt | 72.00 | 100.00 | 100.00 | 91.70 | 70.00 pass | 84.60 | 46.8 | 49.9 | 100.0 | |
| GPT-6 Astra gpt | 67.00 | 89.30 | 100.00 | 87.50 | 90.00 pass | 77.04 | 38.2 | 49.1 | 100.0 | |
| GPT-6.1 Sol gpt | 67.00 | 89.30 | 100.00 | 87.50 | 90.00 pass | 77.04 | 96.5 | 49.1 | 100.0 | |
| Gemini 3.1 Pro gemini | 42.00 | 89.30 | 100.00 | 91.70 | 90.00 pass | 63.29 | 83.3 | 38.3 | 100.0 | |
| GLM-4.6 zhipu | 42.00 | 89.30 | 25.00 | 91.70 | 80.00 pass | 63.29 | 95.0 | 66.0 | 75.0 | |
| Qwen3 Max qwen | - | - | - | - | - | - | 0.0 | 75.0 | 0.0 |
API 访问：如需程序化访问评测数据，请使用我们的 API。