YZ Index
评测数据
当前展示:Run #430 | 2026-11-16 | 128 题库 | 公式 v7 | 判分 v6.5
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 |
|---|---|---|---|---|---|---|---|---|---|---|
| GPT-6 Astra gpt | 88.90 | 69.80 | 89.60 | 83.60 | 83.30 pass | 80.31 | 8.1 | 40.2 | 100.0 | |
| GPT-6 Sol gpt | 89.60 | 67.60 | 89.00 | 86.10 | 90.00 pass | 79.70 | 38.8 | 43.8 | 100.0 | |
| Claude Opus 4.7 claude | 80.30 | 78.60 | 73.60 | 65.80 | 90.00 pass | 79.54 | 5.3 | 33.6 | 100.0 | |
| GPT-6.1 Sol gpt | 88.80 | 65.00 | 89.60 | 86.30 | 83.30 pass | 78.09 | 38.1 | 38.1 | 100.0 | |
| GPT-5.5 gpt | 82.60 | 70.60 | 87.80 | 87.70 | 86.70 pass | 77.20 | 18.9 | 42.6 | 98.0 | |
| GPT-6 Luna gpt | 78.50 | 73.00 | 86.50 | 79.40 | 78.30 pass | 76.03 | 100.0 | 36.1 | 99.0 | |
| Claude Sonnet 4.6 claude | 82.10 | 67.60 | 69.70 | 63.80 | 83.30 pass | 75.58 | 24.3 | 31.7 | 100.0 | |
| Grok 4 grok | 75.10 | 70.80 | 44.10 | 66.50 | 76.70 pass | 73.17 | 22.5 | 28.3 | 100.0 | |
| 豆包 Pro doubao | 82.30 | 60.40 | 88.90 | 87.10 | 90.00 pass | 72.45 | 94.6 | 38.3 | 97.8 | |
| GPT-o3 gpt | 65.50 | 76.40 | 83.40 | 75.80 | 81.70 pass | 70.41 | 8.8 | 37.3 | 91.0 | |
| DeepSeek V4 Pro deepseek | 62.10 | 76.20 | 87.80 | 84.20 | 76.70 pass | 68.45 | 45.8 | 44.5 | 90.8 | |
| GLM-4.6 zhipu | 47.30 | 75.10 | 65.70 | 55.80 | 86.70 pass | 59.81 | 41.4 | 41.1 | 80.7 | |
| Gemini 3.1 Pro gemini | 50.90 | 65.50 | 82.80 | 75.80 | 80.00 pass | 57.47 | 21.5 | 18.3 | 100.0 | |
| Gemini 2.5 Pro gemini | 48.70 | 67.00 | 80.10 | 76.50 | 90.00 pass | 56.94 | 32.7 | 36.4 | 83.0 | |
| 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。