YZ Index
评测数据
当前显示：Run #386 | 2026-10-19 | 154题题库 | 公式 v7 | 判定 v6.4
数据公开说明：为防止题库污染和过拟合，题目原文与期望答案不公开。本页面展示模型回答、评分、判定方式等透明数据。完整方法论请参见方法论页面。
| 模型 | 代码执行 | 资料约束 | 工学判断 | 任务表达 | 誠実性 | 主评分 | 性价比 | 安定性 | 可用性 | 详细操作 |
|---|---|---|---|---|---|---|---|---|---|---|
| GPT-6 Sol gpt | 88.30 | 70.80 | 85.50 | 86.10 | 86.70 pass | 80.43 | 39.0 | 41.8 | 100.0 | |
| Grok 4 grok | 83.30 | 74.80 | 67.00 | 82.90 | 81.70 pass | 79.48 | 25.5 | 33.3 | 99.0 | |
| Claude Opus 4.7 claude | 78.70 | 79.90 | 82.20 | 72.80 | 93.30 pass | 79.24 | 5.4 | 33.4 | 100.0 | |
| GPT-o3 gpt | 79.80 | 76.60 | 82.20 | 78.30 | 76.70 pass | 78.36 | 9.5 | 36.5 | 98.0 | |
| GPT-6.1 Sol gpt | 82.50 | 73.00 | 89.60 | 83.60 | 81.70 pass | 78.23 | 38.2 | 36.5 | 100.0 | |
| Claude Sonnet 4.6 claude | 82.00 | 71.80 | 83.20 | 89.50 | 75.00 pass | 77.41 | 25.8 | 37.2 | 100.0 | |
| GPT-5.5 gpt | 82.50 | 70.60 | 87.80 | 90.80 | 81.70 pass | 77.15 | 18.9 | 37.7 | 100.0 | |
| GPT-6 Luna gpt | 79.60 | 73.80 | 80.70 | 86.30 | 86.70 pass | 76.99 | 100.0 | 35.6 | 100.0 | |
| GPT-6 Astra gpt | 82.50 | 70.00 | 89.60 | 83.60 | 86.70 pass | 76.88 | 8.0 | 36.5 | 100.0 | |
| DeepSeek V4 Pro deepseek | 68.50 | 82.10 | 83.90 | 73.60 | 83.30 pass | 74.62 | 44.7 | 49.6 | 91.4 | |
| 豆包 Pro doubao | 78.10 | 67.60 | 90.00 | 93.30 | 78.30 pass | 73.38 | 94.1 | 38.6 | 97.8 | |
| Gemini 3.1 Pro gemini | 68.90 | 75.40 | 80.00 | 78.30 | 86.70 pass | 71.83 | 24.9 | 27.3 | 100.0 | |
| Gemini 2.5 Pro gemini | 63.60 | 71.50 | 72.90 | 78.30 | 85.00 pass | 67.16 | 35.3 | 32.1 | 89.9 | |
| GLM-4.6 zhipu | 32.70 | 44.80 | 30.00 | 35.00 | 26.70 fail | 38.15 | 21.4 | 48.0 | 40.4 | |
| Qwen3 Max qwen | - | - | - | - | - | - | 0.0 | 75.0 | 0.0 |
API访问：如需通过程序访问评测数据，请使用
API 。。