YZ Index
评测数据
当前展示:Run #361 | 2026-10-05 | 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 |
|---|---|---|---|---|---|---|---|---|---|---|
| Claude Opus 4.7 claude | 94.50 | 97.00 | 94.50 | 53.90 | 100.00 pass | 95.63 | 29.6 | 51.4 | 100.0 | |
| GPT-o3 gpt | 94.50 | 85.30 | 88.90 | 90.00 | 100.00 pass | 90.36 | 49.7 | 68.3 | 100.0 | |
| GPT-6 Sol gpt | 94.50 | 83.10 | 94.80 | 90.00 | 100.00 pass | 89.37 | 97.9 | 71.7 | 100.0 | |
| 豆包 Pro doubao | 91.30 | 85.30 | 100.00 | 95.00 | 90.00 pass | 88.60 | 100.0 | 72.4 | 100.0 | |
| Gemini 2.5 Pro gemini | 91.30 | 85.30 | 75.00 | 95.00 | 100.00 pass | 88.60 | 98.3 | 76.2 | 95.0 | |
| DeepSeek V4 Pro deepseek | 91.30 | 73.10 | 63.90 | 95.00 | 100.00 pass | 83.11 | 98.7 | 67.6 | 95.0 | |
| Qwen3 Max qwen | 65.50 | 94.20 | 61.10 | 63.90 | 100.00 pass | 78.42 | 99.4 | 39.1 | 100.0 | |
| Gemini 3.1 Pro gemini | 69.50 | 85.30 | 75.00 | 85.00 | 100.00 pass | 76.61 | 88.1 | 35.8 | 100.0 | |
| GPT-5.5 gpt | 69.50 | 83.10 | 88.90 | 90.00 | 100.00 pass | 75.62 | 75.4 | 49.2 | 100.0 | |
| GPT-6 Luna gpt | 69.50 | 83.10 | 89.20 | 90.00 | 100.00 pass | 75.62 | 100.0 | 51.2 | 100.0 | |
| GPT-6 Astra gpt | 69.50 | 78.30 | 97.80 | 90.00 | 100.00 pass | 73.46 | 38.5 | 49.7 | 100.0 | |
| Grok 4 grok | 64.00 | 83.10 | 80.60 | 81.10 | 100.00 pass | 72.60 | 86.2 | 42.8 | 100.0 | |
| GPT-6.1 Sol gpt | 69.50 | 73.10 | 97.80 | 85.00 | 100.00 pass | 71.12 | 96.2 | 47.2 | 100.0 | |
| Claude Sonnet 4.6 claude | 44.50 | 88.30 | 100.00 | 45.00 | 100.00 pass | 64.21 | 83.1 | 21.1 | 100.0 | |
| GLM-4.6 zhipu | 44.50 | 38.30 | 75.00 | 25.00 | 90.00 pass | 41.71 | 89.8 | 66.9 | 60.0 |
API 访问:如需程序化访问评测数据,请使用我们的 API。