编程：答案错误
编程
答案错误
看看哪些 AI 模型在 编程 上最容易遇到 答案错误，更快找出薄弱点。
197/197
筛选模型
没有模型匹配当前搜索和筛选条件。
| 排名 | 模型 | 公司 | 答案错误 次数 | 分类得分 | 总成本 | 测试正确 | 响应时间（平均） |
|---|---|---|---|---|---|---|---|
| #46 | Gemini 2.5 Flash medium | 1 | 7.8 | $0.644 | 2/3 | 41.0s | |
| #47 | Claude Sonnet 5 medium | Anthropic | 1 | 9.0 | $0.922 | 2/3 | 17.3s |
| #51 | Inkling high | Thinkingmachines | 1 | 8.5 | $1.041 | 2/3 | 148.4s |
| #53 | DeepSeek V4 Flash 0731 low | DeepSeek | 1 | 7.3 | $0.072 | 1/3 | 163.3s |
| #56 | Inkling medium | Thinkingmachines | 1 | 8.4 | $0.377 | 2/3 | 25.2s |
| #57 | GPT-5.6 Terra high | OpenAI | 1 | 7.6 | $0.836 | 2/3 | 9.14s |
| #59 | GPT-5.2 Chat none | OpenAI | 1 | 8.8 | $0.594 | 2/3 | 9.82s |
| #60 | Step 3.7 Flash medium | Stepfun | 1 | 8.8 | $0.495 | 2/3 | 27.4s |
| #62 | Qwen3.7 Plus medium | Qwen | 1 | 6.1 | $0.277 | 1/3 | 108.6s |
| #63 | Qwen3.6 Plus medium | Qwen | 1 | 6.1 | $0.418 | 1/3 | 153.1s |
| #64 | Gemini 3.5 Flash Lite medium | 1 | 7.9 | $0.272 | 2/3 | 9.96s | |
| #65 | GLM 5.2 medium | Z.ai | 1 | 8.2 | $0.224 | 2/3 | 41.0s |
| #67 | Claude Sonnet 4.6 medium | Anthropic | 1 | 5.7 | $2.126 | 1/3 | 33.3s |
| #69 | Seed-2.0-Lite medium | Bytedance Seed | 1 | 8.0 | $0.236 | 2/3 | 156.7s |
| #70 | Claude Opus 4.8 low | Anthropic | 1 | 6.6 | $2.089 | 1/3 | 7.58s |